research papers on artificial intelligence and ipr

• Search Menu
• Advance articles
• Author Guidelines
• Submission Site
• Open Access Options
• About Journal of Intellectual Property Law & Practice
• Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Self-Archiving Policy
• Dispatch Dates
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

1. introduction, 2. computer-generated works, 3. text and data mining (tdm), 4. patent inventorship.

• < Previous

Artificial intelligence and intellectual property: copyright and patents—a response by the CREATe Centre to the UK Intellectual Property Office’s open consultation

• Article contents
• Figures & tables
• Supplementary Data

Martin Kretschmer, Bartolomeo Meletti, Luis H Porangaba, Artificial intelligence and intellectual property: copyright and patents—a response by the CREATe Centre to the UK Intellectual Property Office’s open consultation, Journal of Intellectual Property Law & Practice , Volume 17, Issue 3, March 2022, Pages 321–326, https://doi.org/10.1093/jiplp/jpac013

• Permissions Icon Permissions

Martin Kretschmer is Professor of Intellectual Property Law; Bartolomeo Meletti is Creative Director; Luis H Porangaba is Lecturer in Intellectual Property Law—all at CREATe, School of Law, University of Glasgow. The authors are named in alphabetical order, contributed equally and should all be cited.

This article

The UK Intellectual Property Office consulted between October 2021 and January 2022 on policy options for intellectual property (IP) law interventions that could ‘secure the UK’s position amongst the global AI superpowers’, in line with the government’s national AI Strategy (September 2021) and the vision ‘to make the UK a global hub for innovation by 2035’ (UK Innovation Strategy, July 2021). This article reproduces the submission by the Copyright and Creative Economy (CREATe) Centre at the University of Glasgow. We show that policymakers are in a difficult position to assess reform proposals relating to artificial intelligence (AI) and IP because evidence remains scarce.

With respect to computer-generated works and patent inventorship, we urge caution. There is no evidence that new rights are needed. The onus of proof needs to lie with the proponents of much discussed proposals, such as offering AI copyright authorship in the guise of computer-generated works or granting AI inventorship under patent law. With respect to text and data mining (TDM), we see a straightforward opportunity to stimulate UK innovation and improve the transparency of AI systems by opening up the current ‘Hargreaves’ exception to all users (Copyright, Designs and Patents Act 1988: s 29A, Copies for text and data analysis for non-commercial research).

More generally, the UK’s research and innovation environment would in our view benefit from a technologically neutral, open-ended user exception (akin to the copyright doctrine of ‘fair use’ in the USA).

‘Artificial intelligence (AI) is a transformative technology, which is already revolutionising many areas of our lives. Unleashing the power of AI is a top priority in the plan to be the most pro-tech government ever’. Thus opens modestly the consultation on Artificial Intelligence and Intellectual Property: Copyright and Patents , conducted by the UK Intellectual Property Office (IPO) between 29 October 2021 and 7 January 2022. 1

The Consultation sought ‘evidence and views’ on three specific areas:

– Copyright protection for computer-generated works without a human author. These are currently protected in the UK for 50 years. But should they be protected at all and if so, how should they be protected?

– Licensing or exceptions to copyright for text and data mining (TDM), which is often significant in AI use and development.

– Patent protection for AI-devised inventions. Should we protect them and if so, how should they be protected?

We consider each of these legal issues in turn, reproducing the given policy options at the beginning of each section, even where we would have preferred to frame the discussion differently. We then proceed to assess the existing evidence. The text is an authentic reproduction of CREATe’s submission to the Consultation.

In the structured format of the Consultation, there was no space to evaluate fully the evidence for our preferred policy option of a technologically neutral, open-ended user exception (‘fair use’). 2

Policy options offered by the UK IPO

With no counterpart in most jurisdictions, s 9(3) of the Copyright, Designs and Patents Act 1988 (CDPA) is rather unique, if not problematic. Indeed, the effective operation of this provision may depend upon other aspects of copyright law which, following Brexit, remain unsettled. By providing 50-year protection to ‘authorless’ computer-generated literary, dramatic, musical or artistic (LDMA) works, s 9(3) poses the complex legal question of what originality standard should be applied. There is an apparent inconsistency with the EU standard of ‘an author’s own intellectual creation’, which relies on creative choices made by an individual, 3 for example. The standard of ‘originality’ applicable to computer-generated outputs that do not reflect human creative input is a matter for UK law alone. 4

In more than 30 years, s 9(3) was only ever considered in a single court decision, 5 which did not address the originality issue. Determining the author of computer-generated works—that is, the ‘person by whom the arrangements necessary for the creation of the work are undertaken’—is no straightforward matter either. In Nova Productions , the Court of Appeal found such a person to be the author of the computer program rather than the user. However, this decision concerned a simple two-dimensional video game, offering limited guidance on the issue of AI-assisted outputs. Furthermore, as the experience with other types of subject matter (eg sound recordings) suggests, the notion of ‘arrangements necessary’ is not resolved, nor is it clear if the ‘person’ making such arrangements can be a legal entity (ie a firm). 6

The introduction of a related right of reduced scope and duration referred to as option 2 may lead to an issue of cumulation, with the same subject matter attracting rights of different kind, as the recent experience with databases suggests. The potential costs of additional IP rights typically are of two kinds: higher prices and loss of innovation. In the UK, the Hargreaves (2011) and Gower (2006) Reviews recommended making the policy process more transparent and rigorous. 7 IP rights, once created, have proved almost impossible to remove. 8 In a period of rapid technological and industrial change, the standards of evidence required therefore must be particularly high. A fundamental point relates to the onus of proof. Advocates of new rights need to evidence what the potential costs are, who will carry them and that the costs are necessary and proportionate to the claimed benefits.

The UK government should carefully consider whether the complex legal questions attendant on AI outputs must really be addressed at such an early stage, as an attempt to anticipate issues that have not emerged yet. There is no conclusive evidence showing that the current copyright framework provides suboptimal incentives for the creation of AI-generated works, let alone the existence or sudden emergence of market failure requiring legislative intervention. The UK IPO’s Impact Assessment is framed by a utilitarian discourse which, if unaccompanied by market-based evidence, may seem all too speculative.

At present, the role of most AI tools is largely limited to the execution stage of creative production, with human authors retaining control over the conception and redaction phases. 9 From a creative and legal perspective, AI applications such as Grammarly are not very different from editing or motion graphics software such as Adobe Photoshop or After Effects. In all such cases, the computer (or AI) carries out the work under the instruction and control of a (human) creator. UK copyright already affords protection to outputs generated by or through such applications so long as they fall within one of the categories of protected works and meet the originality standard. 10

There is no real need for a dedicated, sui generis provision dealing with copyright subsistence in computer-generated works. Unless strong evidence emerges that AI users, developers and businesses indeed do rely on s 9(3), we recommend that the UK government removes protection for computer-generated works (Option 1) .

Extracting information from copyright-protected materials should not be considered a copyright-relevant act. We, therefore, recommend that the UK should avail herself of recently acquired post-Brexit freedoms to foster innovation by adopting a TDM exception for any use (Option 4) . In addition, the introduction of a technologically neutral, open-ended exception (akin to the fair use doctrine in the USA) should be explored.

With regard to TDM, more evidence is available than for the issue of computer-generated works. Indeed, empirical research indicates that in jurisdictions with more permissive copyright frameworks 11 and robust research exceptions, more data mining-related research is conducted. 12 Higher firm revenues in information industries, computer system design and software publishing as well as increased, higher-quality scholarly output appear to be found in countries with more open user-friendly provisions such as the US fair use clause. 13 The scope of the UK exception for text and data mining (s 29A CDPA) is rather narrow and uncertain, creating confusion, for example in the context of the widespread practice of data scraping. 14 While Option 4 seems the most conducive to innovation in research and business, we would favour two other options which the UK government may not have considered: (i) excluding from the scope of exclusive rights TDM and other acts of extracting informational value from protected works; or (ii) introducing a technologically neutral, open-ended exception akin to the fair use doctrine in the USA.

We understand that option (i) could be effected by judicial interpretation, especially in a post-Brexit context allowing UK courts to depart from the jurisprudence of the Court of Justice of the European Union (CJEU). Lord Hoffmann’s speech in Designer’s Guild , for example, makes it clear that copyright protection should not extend to the ideas underlying LDMA works. 15 If ‘a literary work which describes a system or invention does not entitle the author to claim protection for his system or invention as such,’ the same equally applies to text and data mining which are more concerned with accessing the information disclosed in—rather than taking the expression of—protected works. 16

Introducing an open-ended exception (ii), however, is a matter of legislation. In 2011, Professor Hargreaves was specifically asked to investigate the benefits of fair use and how it could be implemented in the UK. 17 At the time, the Hargreaves Review concluded that the introduction of fair use into UK law would likely be inconsistent with the EU copyright framework. Instead, the Review recommended the adoption of several closed exceptions stemming from the InfoSoc Directive, including text and data mining. Following Brexit, now may be the time for the UK to rethink fair use.

Innovation is determined by a wide variety of economic, cultural, political and social factors, and in the field of copyright, fair use has been a successful legal mechanism in promoting it. In the USA, fair use has allowed the emergence of indexing and search technology, the Google Books project, and, more recently, the copying of code from the Java API into the Android operating system. 18 The recent jurisprudence of the US Supreme Court may suggest that most AI-related uses of copyright works are likely to fall within fair use. Would these types of innovation and other potential applications of AI be equally accommodated by rule-based, purpose-limited exceptions such as copying for text and data analysis? We do not think so.

The interests of rightholders are of course legitimate. However, the proposed Option 1 of developing a licensing environment that would provide lawful access to the underlying data within countless copyright works seems unrealistic. Requiring rights clearance for TDM and other AI uses of protected materials would increase transaction costs significantly, raising entry barriers for small and medium-sized enterprises (SMEs), in particular market entrants. Big tech corporations would likely retain access to enormous, high-quality, exponentially growing amounts of data, while others ‘may find it economically attractive to train their algorithms on “cheaper”, which often means older, less accurate or biased, data’. 19

Based on the evidence currently available (or the lack thereof), we argue that no reform is necessary in this area (Option 0) .

Significantly, there is no compelling economic evidence or policy for AI to be formally recognized as ‘inventor’. Unlike (real) human inventors, AI does not have a moral claim to inventorship, neither do we anticipate any disputes relating to entitlement to grant to arise from a purported ‘AI inventor’. We share the view (and frustration) of other legal scholars 20 and practitioners 21 that the AI inventorship debate is seriously overblown and, indeed, seems to be detracting from other, more significant issues in the field. The existing patent framework is fully capable to accommodate technological developments in AI, just as it has been done with biotechnology. 22

Furthermore, any reforms which the government may understand to be required should be implemented at the international level, which may not seem achievable or even realistic at this juncture. Formal recognition of this putative inventorship would have to be mirrored across most patent systems; otherwise, applications claiming UK priority may be found incompatible, raising significant barriers to and associated costs with international prosecution. The European Patent Office (EPO), for example, has recently confirmed on appeal the rejection of the DABUS applications EP 18 275 163 and EP 18 275 174. While the decision has yet to be made publicly available, the EPO made it clear that ‘only a human inventor could be an inventor’ and ‘a machine could not transfer any rights to the applicant.’ 23 Unless harmonization is sought—and hopefully achieved—via international law, interventions at the national level will only risk inconsistency. Rather, the current UK position, following the Court of Appeal’s judgement in Thaler v Comptroller General of Patents , 24 is one of relative legal certainty—we do know that AI cannot be named as inventor and, empirically, nothing suggests there is a pressing need for this to be changed.

Hence, the reform proposals 1 and 2 under consideration would run counter to the tradition of UK patent law which has been largely developed by judicial practice striving for consistency with EPO decisions. 25 In a rapid developing field such as AI, ex post regulation through minor doctrinal adjustments within the discretion of courts and patent offices should be the norm. In the past, more significant policy issues such as the patentability of second medical use inventions have been addressed this way under the European Patent Convention. 26 Legislative intervention of the kind being proposed is unwarranted, running the risk of increasing transaction costs associated with patent protection without any tangible benefit.

Particularly, there is no conclusive evidence that AI systems can effectively invent autonomously. 27 Indeed, the previous call for views on AI concluded that ‘there appeared to be near complete agreement that AI systems are not, or not yet, independent agents seeking patent rights without human intervention’. 28 In response to that consultation, IBM stated that ‘AI with the ability to invent without the assistance of a human is a considerable way off … We believe that AI will remain tools that assist humans, rather than invent independently and autonomously, for a considerable time’. 29 It is therefore not surprising that some have questioned the ability and legitimacy of the so-called DABUS system, which is not sufficiently explained in any of the patent applications referencing it. Put this way, one might speculate whether the Thaler litigation amounts to anything other than a publicity stunt. 30

The introduction of a new, sui generis right to protect AI-devised inventions referred to as Option 2 would also be ill-advised. This would significantly increase costs associated with determining the content of this law, including matters of prosecution and enforcement, which may have a differential impact on small and medium-sized enterprises (SMEs) in the field of technology. There is no guarantee that this new form of protection would develop in the same way as or even build on the existing patent jurisprudence, for example. Recent experience with database rights and supplementary protection certificates both illustrate the difficulty in determining, let alone predicting, how the relevant statutory provisions will be interpreted and applied.

By and large, patent applications for AI-related inventions—particularly those featuring deep learning and neural networks—are expected to increase over the next years. 31 Even if AI reaches the stage of developing inventions with minimal or no human intervention and those outputs prove to be unpatentable on such grounds, there is no economic evidence indicating this would be detrimental to the investment in and the development of AI technology. As a practitioner has suggested, ‘the main commercial players in the AI field, such as Google DeepMind, continue to navigate the patent system without apparent concern about the issue of AI inventorship’. 32

UK Intellectual Property Office, Artificial Intelligence and Intellectual Property: Copyright and Patents (29 October 2021). Available at https://www.gov.uk/government/consultations/artificial-intelligence-and-ip-copyright-and-patents/artificial-intelligence-and-intellectual-property-copyright-and-patents (accessed March 1 2022).

We understand that the UK government is considering the matter but may treat it as too politically charged for an open consultation at this stage. For the record, we consider a well-conceived ‘fair use’ exception to be an author-friendly provision compatible with measures relating to remuneration and contracts (which we have supported in different contexts). Cf Séverine Dusollier et al. ‘Comment of the European Copyright Society Addressing Selected Aspects of the Implementation of Articles 18 to 22 of the Directive (EU) 2019/790 on Copyright in the Digital Single Market’ (2020) 11 Journal of Intellectual Property, Information Technology and Electronic Commerce Law 133.

See for example Infopaq International A/S v Danske Dagblades Forening , C-5/08, EU:C:2009:465; Football Association Premier League Ltd v QC Leisure , C-403/08 and 429/08, EU:C:2011:631; SAS Institute Inc v World Programming Ltd , C-406/10, EU:C:2012:259.

Lionel Bently et al. Intellectual Property Law (5th edn Oxford University Press 2018) 118.

Nova Productions Ltd v Mazooma Games Ltd [2007] EWCA Civ 219.

CDPA s 9(2) defines the producer as the author of a sound recording. Under CDPA s 178, ‘producer’, in relation to a sound recording or a film, means the person ‘by whom the arrangements necessary’ for the making of the sound recording or film are undertaken. Case law indicates that financial and organizational inputs are important ( Beggars Banquet Records v Carlton TV [1993] EMLR 349). This can range from access to venues to contracting performers. In Bamgboye v Reed [2004] EMLR 5 para 85, Hazel Williamson QC (sitting as a Deputy Judge of the High Court) said: ‘It seems to me the question can be summarised in this sense: Looking at the cases, and remembering that it is always a question of fact. Who was it who got the recording made (to put it in a colloquial way)?’ Unfortunately, no clearer notion of what are ‘arrangements necessary’ has emerged since.

Andrew Gowers, Gowers Review of Intellectual Property (2006); Ian Hargreaves, Digital Opportunity: A Review of Intellectual Property and Growth (2011). Recommendation 1 of the Hargreaves Report reads: ‘Government should ensure that development of the IP system is driven as far as possible by objective evidence. Policy should balance measurable economic objectives against social goals and potential benefits for rights holders against impacts on consumers and other interests. These concerns will be of particular importance in assessing future claims to extend rights or in determining desirable limits to rights’.

Martin Husovec, ‘The Fundamental Right to Property and the Protection of Investment: How Difficult Is It to Repeal New Intellectual Property Rights?’ in Christophe Geiger (ed) Research Handbook on Intellectual Property and Investment Law (Edward Elgar 2020) 385–405.

P Bernt Hugenholtz and João Pedro Quintais, ‘Copyright and Artificial Creation: Does EU Copyright Law Protect AI-Assisted Output?’ (2021) 52 International Review of Intellectual Property and Competition Law 1190.

While there are some AI systems—such as the GP-T2 and GP-T3 text generators developed by OpenAI—that can generate creative outputs with minimal contribution from individuals, this does not change the analysis. For an example of GP-T3, see Opinion Artificial Intelligence, ‘A Robot Wrote This Entire Article. Are You Scared Yet, Human?’ The Guardian (8 September 2020) https://www.theguardian.com/commentisfree/2020/sep/08/robot-wrote-this-article-gpt-3 accessed 23 December 2021. We agree with Goold that s 9(3) ‘is either unnecessary or unjustifiably extends legal protection to a class of works which belong in the public domain’. See Patrick Goold, ‘The Curious Case of Computer-Generated Works under the Copyright, Designs and Patents Act 1988’ [2021] Intellectual Property Quarterly 120.

Christian Handke et al . ‘Copyright’s Impact on Data Mining in Academic Research’ (2021) 42 Managerial Decision Economics 1.

Michael Palmedo, ‘The Impact of Copyright Exceptions for Researchers on Scholarly Output’ (2019) 2 Efil Journal of Economic Research 114.

Sean Flynn and Michael Palmedo, ‘The User Rights Database: Measuring the Impact of Copyright Balance’ (2019) Digital Commons @ American University Washington College of Law Working Papers 42. Available at https://digitalcommons.wcl.american.edu/fac_works_papers/42 (accessed March 1 2022).

Sheona Burrow, ‘The Law of Data Scraping: A Review of UK Law on Text and Data Mining’ (March 2021) CREATe Working Paper 2021/2. Available at https://www.create.ac.uk/blog/2021/03/30/new-working-paper-the-law-of-data-scraping-a-review-of-uk-law-on-text-and-data-mining/ (accessed March 1 2022).

Designers Guild Ltd v Russell Williams (Textiles) Ltd [2000] 1 WLR 2416, HL, 2423 (‘… a copyright work may express certain ideas which are not protected because they have no connection with the literary, dramatic, musical or artistic nature of the work. It is on this ground that, for example, a literary work which describes a system or invention does not entitle the author to claim protection for his system or invention as such. The same is true of an inventive concept expressed in an artistic work.’). See also Newspaper Licensing Agency Ltd v Marks & Spencer Plc [2001] UKHL 38, [19]–[27].

See also Catnic Components Limited v Hill and Smith Limited [1982] RPC 183, HL, 223, rejecting infringement of artistic copyright based on the taking of drawings of lintels used in construction (‘If an ‘artistic work’ is designed to convey information, the importance of some part of it may fall to be judged by how far it contributes to conveying that information, but not, in my opinion, by how important the information may be which it conveys or helps to convey. What is protected is the skill and labour devoted to making the ‘artistic work’ itself, not the skill and labour devoted to developing some idea or invention communicated or depicted by the ‘artistic work’. The protection afforded by copyright is not, in my judgment, any broader, as counsel submitted, where the ‘artistic work’ embodies a novel or inventive idea than it is where it represents a commonplace object or theme’).

Hargreaves (n 8).

Authors Guild v Google Inc 770 F Supp 2d (SDNY 2011); Authors Guild v Google Inc 721 F3d (2nd Cir 2013); Authors Guild v Google Inc 954 F Supp 2d (SDNY 2013); Authors Guild v Google Inc 804 F3d (2nd Cir 2015); Google LLC v Oracle America Inc 141 S Ct 1163 (2021).

The narrow current UK exception may in fact incentivise the import of AI models already trained on unverifiable data. See Thomas Margoni and Martin Kretschmer, ‘A Deeper Look into the EU Text and Data Mining Exceptions: Harmonisation, Data Ownership, and the Future of Technology’ (July 2021) CREATe Working Paper 2021/7. Available at https://www.create.ac.uk/blog/2021/07/14/ai-machine-learning-and-eu-copyright-law/ (accessed March 1 2022).

Dan L Burk, ‘AI Patents and the Self-Assembling Machine’ in Daniel J Gervais (ed) The Future of Intellectual Property (Edward Elgar 2021).

Rose Hughes, ‘DABUS: An AI inventor or the Emperor’s New Clothes?’ ( IPKat Blog , 15 September 2021). https://ipkitten.blogspot.com/2021/09/dabus-ai-inventor-or-emperors-new.html accessed 23 December 2021.

See, in particular, Burk (n 21) 130 (‘… far from challenging the existing order of patent law, the patent system is fully equipped to encompass AI innovation, with perhaps some minor doctrinal accommodations that are well within the policy lever discretion available to the courts and to the patent offices.’). While the EU enacted the Biotechnology Directive, the jurisprudence and practice of the European Patent Office largely draws on the general provisions of the European Patent Convention.

European Patent Office, Press Communiqu é on Decisions J 8/20 and J 9/20 of the Legal Board of Appeal (21 December 2021) https://www.epo.org/law-practice/case-law-appeals/communications/2021/20211221.html accessed 23 December 2021.

Thaler v Comptroller General of Patents, Trade Marks and Designs [2021] EWCA Civ 1374.

See also Merrell Dow Pharmaceuticals Inc v Norton & Co Ltd [1995] UKHL 14, [12] (‘[UK courts] must have regard to the decisions of the European Patent Office (“EPO”) on the construction of the EPC. These decisions are not strictly binding upon courts in the U.K. but they are of great persuasive authority; first, because they are decisions of expert courts (the Boards of Appeal and Enlarged Board of Appeal of the EPO) involved daily in the administration of the EPC and secondly, because it would be highly undesirable for the provisions of the EPC to be construed differently in the EPO from the way they are interpreted in the national courts of a Contracting State’).

See, in particular, Eisai/Second Medical Indication G05/83 [1979–85] EPOR B241 (Enlarged Board of Appeal); John Wyeth & Brother Ltd’s Application [1985] RPC 545. For an account of the judicial development of Swiss-form claims, see also Actavis UK Ltd v Merck & Co Inc [2008] EWCA Civ 444, [7]–[48].

Daria Kim et al. ‘Artificial Intelligence Systems as Inventors? A Position Statement of 7 September 2021 in View of the Evolving Case Law Worldwide’ (September 2021) Max Planck Institute for Innovation and Competition, 5. Available at https://www.ip.mpg.de/fileadmin/ipmpg/content/stellungnahmen/MPI_Position_statement_AI_Inventor_2021-08-09.pdf (accessed March 1 2022). See also Burk (n 21) 131 (‘Such systems are not intelligent in any robust sense of the word; they lack any hint or expectation of encompassing “strong” AI with general cognitive abilities of the sort that humans (or even animals) routinely display. There is at present no serious prospect of designing machines with such capabilities …’).

Government Response to Call for Views on Artificial Intelligence and Intellectual Property (23 Mar 2021).

IBM response (20 November 2020) https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/971519/Response-to-AI-5.zip accessed 29 December 2021.

See also Rose Hughes, ‘The First AI Inventor—IPKat Searches for the Facts Behind the Hype’ ( IPKat Blog , 15 August 2019) https://ipkitten.blogspot.com/2019/08/the-first-ai-inventor-ipkat-searches.html accessed 29 December 2021.

World Intellectual Property Organization, WIPO Technology Trends 2019: Artificial Intelligence (2019).

Hughes (n 22).

Email alerts

Citing articles via.

• About Journal of Intellectual Property Law & Practice
• JIPLP Weblog
• Recommend to your Library

Affiliations

• Online ISSN 1747-1540
• Print ISSN 1747-1532
• Copyright © 2023 Oxford University Press
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2023 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

{{locationData.PopupWindowHeading}}

• {{ !!location.countrycode?location.countryName :location.officeName }}
• {{lL.name }} {{lL.languageCode | uppercase}}
• Skip to main navigation
• Skip to content
• Skip to footer

0"> {{suggestionHead.categoryName}}

{{bckdata.locationheading}}, {{headerdata.hamburgerprimaryfeatureheading}}, {{tile.title}}, {{headerdata.hamburgersecondaryfeatureheading}}, artificial intelligence and ip rights: threats and opportunities.

The EUIPO study

On March 2, 2022 the European Union Intellectual Property Office (EUIPO) published its “Study on the impact of artificial intelligence on the infringement and enforcement of copyright and designs” (the Study), describing current and future connections between IP law and artificial intelligence and the implications of technology transformation on the IP system.

The Study was promoted and carried out by the “Impact of Technology Expert Group,” which was set up in 2019 by the EUIPO to better assess the impact of upcoming technology on the infringement and enforcement of intellectual property rights: The idea behind the group’s approach — the so-called “double-edged sword”— is that all emerging and disruptive technologies have the potential to be used as means to enhance IP protection, but they can also be tools for IP infringement.

Building on this principle, the Study illustrates the threats and opportunities raised by AI technologies. It sets out several scenarios for the two storylines of physical products and digital content — for example, the use by counterfeiters of machine learning systems to recognize safe ports where their goods are more likely to pass through customs authorities, or the use of AI-supported blockchain by IP offices to protect information in registration systems.

These hypothetical cases are significant examples of the twofold nature of AI from an IP perspective.

Practical scenarios: infringement and enforcement of IP rights on content-sharing platforms

The Study explains how AI technologies can be used in different cases involving the upload of protected materials online, which is of particular interest also in view of EU Directive No. 790 of 2019 (the Copyright Directive, already implemented in Italy and in 15 other EU member states) and, specifically, of the relevant provisions on the liability of online content-sharing service providers for infringing user-generated content.

For example, machine learning tools can be deployed to remove the digital dots and watermarks used to track the distribution of unauthorized copies of copyright works online, as well as to generate “deep fakes”—defined within the Study as “synthetic media in which an individual in an image or video is replaced with another's likeness”—which are developed through a specific type of machine learning, known as “generative adversarial networks”. Based on models that learn how to produce new data with the same characteristics of the “training” data that was fed into the network, artificial images, videos or sounds can be created resembling the natural and/or original ones.

On the other hand, authorities can harness that same technology in a positive way. Once the deep fakes have been identified, the authorities can use AI bots to identify on social media those components of the deep fake that constitute copyright and design infringement.

Also computer vision applications can be implemented in the field of IP, for instance, computer vision applications can determine if videos are original or artificially generated.

The EUIPO also emphasizes the relevance of natural language processing technologies –– these are tools that computers can use to process data in order to understand the meaning of human language. Such technologies can be used by law enforcers to analyze users’ behavior and the content they share, finding correlations in datasets in order to prevent potential future infringements as well as to prove the origin of the counterfeited content.

The new enforcement instruments offered by the copyright directive

AI technologies can also surely be used to enhance the detection — and possibly even prediction — of IP violation. However, the same are increasingly being used for infringing purposes as well. Such uses of AI technologies will be more and more relevant also in the framework of the new discipline introduced by the Copyright Directive.

Under the new regime, user-generated content sharing platforms must obtain the right holders’ authorization when they grant the public access to copyright-protected content uploaded by their users. Lacking this authorization, the platform operators will be held liable for copyright violation, unless they manage to prove that: (i) they have made their best efforts to obtain authorization; (ii) they have made their best efforts to ensure the unavailability of works for which the right holders have provided the necessary information; and (iii) they have acted expeditiously, upon receiving a sufficiently substantiated notice from the right holders, to disable access to, or to remove infringing content, and made their best efforts to prevent future uploads of the same.

On one hand, the deployment of AI can tangibly help service providers in identifying infringement materials, lowering potential liabilities and reinforcing the protection for right holders. On the other hand, the massive use of such technologies for infringing purposes can hinder the detection of possible violations, with increased risks of liability for providers.

Conclusively, several benefits and drawbacks coexist in the increasing use of AI technologies from an IP perspective 1 . In the context of online content-sharing platforms, this calls for a wider use of cutting-edge AI technologies by service providers to tackle infringing activities. Indeed, online content-sharing service providers are not subject to a general surveillance obligation under the Copyright Directive, but they must make best efforts to react expeditiously under certain circumstances.

AI is a “double-edged sword.” Service providers must make sure to have that sword and use it properly to comply with best-efforts requirements in defeating sophisticated AI-based infringements of IP rights.

Carlo Scotto di Clemente  and Camilla Rosi  co-authored this article.

• During Dentons’ AI Trilogy webinar, held on March 31, 2022, Professor Andrea Renda compared artificial intelligence to products considered to be dangerous. As he pointed out: “providing consumers or users of the AI systems with the information needed to understand what are the limitations and the capabilities and the intended use of these AI systems is not very different from many other products that are considered to be dangerous […] the idea is that when we use a pharmaceutical product we have the information on what that product is intended to do and how to use it with care. Handle it with care, not with fear, but with care at least”. Don’t miss the webinar recorded video at the following link: Dentons' AI trilogy webinar series - Developing your AI strategy: the key areas you need to consider - YouTube .

How can the world's largest global law firm help you today ?

Contact us or find an office in your location..

Unsolicited emails and other information sent to Dentons will not be considered confidential, may be disclosed to others, may not receive a response, and do not create a lawyer-client relationship. If you are not already a client of Dentons, please do not send us any confidential information.

Redirection

You are switching to another language. Please click Confirm below to continue.

Leaving Site

You will now be taken from the global Dentons website to the $redirectingsite website. To proceed, please click Accept.

Leaving Dentons

Beijing Dacheng Law Offices, LLP ("大成") is an independent law firm, and not a member or affiliate of Dentons. 大成 is a partnership law firm organized under the laws of the People’s Republic of China, and is Dentons' Preferred Law Firm in China, with offices in more than 40 locations throughout China. Dentons Group (a Swiss Verein) ("Dentons") is a separate international law firm with members and affiliates in more than 160 locations around the world, including Hong Kong SAR, China. For more information, please see dacheng.com/legal-notices or dentons.com/legal-notices.

Important Notice

Innovations in intellectual property rights management: Their potential benefits and limitations

European Journal of Management and Business Economics

ISSN : 2444-8494

Article publication date: 9 April 2019

Issue publication date: 16 July 2019

The purpose of this paper is to evaluate innovations in intellectual property rights (IPR) databases, techniques and software tools, with an emphasis on selected new developments and their contribution towards achieving advantages for IPR management (IPRM) and wider social benefits. Several industry buzzwords are addressed, such as IPR-linked open data (IPR LOD) databases, blockchain and IPR-related techniques, acknowledged for their contribution in moving towards artificial intelligence (AI) in IPRM.

Design/methodology/approach

The evaluation, following an original framework developed by the authors, is based on a literature review, web analysis and interviews carried out with some of the top experts from IPR-savvy multinational companies.

The paper presents the patent databases landscape, classifying patent offices according to the format of data provided and depicting the state-of-art in the IPR LOD. An examination of existing IPR tools shows that they are not yet fully developed, with limited usability for IPRM. After reviewing the techniques, it is clear that the current state-of-the-art is insufficient to fully address AI in IPR. Uses of blockchain in IPR show that they are yet to be fully exploited on a larger scale.

Originality/value

A critical analysis of IPR tools, techniques and blockchain allows for the state-of-art to be assessed, and for their current and potential value with regard to the development of the economy and wider society to be considered. The paper also provides a novel classification of patent offices and an original IPR-linked open data landscape.

• Artificial intelligence
• Software tools
• Social benefits
• Intellectual property rights management
• Linked open databases

Modic, D. , Hafner, A. , Damij, N. and Cehovin Zajc, L. (2019), "Innovations in intellectual property rights management: Their potential benefits and limitations", European Journal of Management and Business Economics , Vol. 28 No. 2, pp. 189-203. https://doi.org/10.1108/EJMBE-12-2018-0139

Emerald Publishing Limited

Copyright © 2019, Dolores Modic, Ana Hafner, Nadja Damij and Luka Cehovin Zajc

Published in European Journal of Management and Business Economics . Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode

1. Introduction

The world today seems to be characterised by the effects of information and communication technology (ICT) on every aspect of our lives, including that of intellectual property rights (IPR) ( Modic, 2017 ). Freeman and Louca (2002 , p. 301) wrote that “even those who have disputed the revolutionary character of earlier waves of technological change, have little difficulty accepting that a vast technological revolution is now taking place”. The surge of intellectual property is mirrored in rising IPR numbers with dissemination efforts dependent upon the available data, channels and skills. IPR data are big data, as its characteristics are high volume, high variety and high velocity of changes ( Ciccatelli, 2017 ). Consequently, merging different types of IPR data from various databases presents a challenge ( Stading, 2017 ; Abbas et al. , 2014 ).

When huge amounts of IPR data are connected, a new ecosystem for (open) innovation emerges. It is important to examine the best available IPR data sources, and their merge-readiness, in order to extract the maximum value. Furthermore, it is important to ensure the availability of appropriate IPR techniques and tools if we are to harness the benefits for IPR management (IPRM) and the wider social benefits of this new open IPR landscape and move towards knowledge creation assisted by artificial intelligence (AI). Examining the latest trends in technological solutions and their potential is the foci of our paper.

Figure 1 presents two dimensions: the benefits and the technology. Looking at the technology dimension, all three layers represent issues companies face. IPR software tools and techniques should better respond to business requirements, and as such support changes in databases when dealing with IPR big data, such as the implementation of blockchain technology and linked open databases.

The benefits dimension is also facing several gaps. One refers to the identification of the accessibility of employees’ knowledge both in SMEs and IPR-savvy companies. In addition, there are inefficiencies when trying to transform tacit to explicit knowledge in order to further knowledge creation.

Both the technology and benefits dimensions are linked, as the technology aims to, largely unsuccessfully at the present time, to support the requirements of the IPRM, thus increasing the IPRM-derived benefits. These would consequently be translated, especially through the use of blockchain technology and IPR-linked open data (IPR LOD) databases, into increased social benefits. The question as to when, and if, the technology will become smart enough to create IPR software tools and techniques that will function in an intelligent manner remains open to debate, as we are faced with increasing transparency and inherently imbued trust.

If AI systems provide the best possible answer to every IPR-related business requirement, in order to maximise business potential, does this mean that the employees’ knowledge creation will become obsolete and AI systems will be able to effectively create new knowledge?

The paper offers a review and an interview-based analysis of the requirements and expectations of some of the top IPR experts from IPR-savvy multinationals, as well as a consideration of the potential social benefits. This is followed by a web-based analysis and data retrieval-based evaluation of the current evolution of IPR (LOD) databases. Furthermore, the practical solutions available have been critically evaluated with respect to IPR databases and IPR software tools. The results of the analysis of the state-of-the-art with the available techniques are presented. Finally, a debate-style conclusion is presented.

2. Background and prepositions

This paper investigates IPRM and IPR social benefits by answering what are the potential social and IPRM benefits of adopting new ICT solutions when dealing with IPR, and especially what is the current state of all three technological layers? The research is based on the following prepositions constructed following the literature review and the evidence-based approach.

The IPR-linked open data (IPR LOD) map is still in its infancy, thus the full potential of their social benefits are still not realized.

AI is a term used very broadly when connected to IPR techniques, to oversell various information retrieval (IR) and machine learning (ML) methods.

The tools do not correspond to the needs of users as expressed by top IPR managers.

Blockchain has the potential to produce both IPRM and IPR-connected social benefits if some issues are solved.

The outputs of this paper are the classifications of IPR databases and patent offices according to Berners-Lee Open Data Plan, and IPR LOD map as connected to patents as well as classification of tools and techniques. A mixed methods approach has been used, every part diligently designed with methodological notes.

3. Methodology

We derive our analysis of potential benefits of new solutions for IPR and the potential of IPR tools from interviews with ten prominent IP experts. First, interviews with ten prominent IP experts were conducted. Seven out of the ten IP experts were head IP managers within their respective companies. The companies selected are positioned highly in terms of patent applications and quality rankings. Furthermore, they appear on top innovation listings, such as MIT’s list of the 50 Smartest companies. All respondents are executives with years of experience; and one of the interviewees appeared twice in the 50 most influential people in IP, as listed by the Managing Intellectual Property magazine. Views expressed inside the interviews are their own and not the views of the companies they are affiliated with. Interviews were conducted either in person, via Skype or via similar VoIP during 2016 and with follow-ups in 2017. Transcripts were analysed using MAXQDA Analytics Pro 12 software. Interview questions were divided into three sections: IPRM (1), formalization (2) and optimisation of processes and gaps reduction (3)). In particular for this paper three topics and their related questions that were included in this semi-structured interview questionnaire are harnessed upon (pertaining to either part (1) or part (3): What is the missing information and/or resources?; Which software tools do you use inside your processes? What are their pros and cons?; What kind of (big) data analysis would be particularly interesting? Who can provide them?

The technologies section brings further methods. The classification of patent offices was done in the period January–February 2018 by conducting web searches and experimental searches with consequent search retrievals inside patent search machines either for full patent documents or at least bibliographical exports. The classification encompasses primarily EU Patent Offices as well as a selection of other relevant patent offices [1] . The framework for the patent map relies on The Linking Open Data cloud diagram, however, it has been significantly upgraded by including material gathered via web searches guided by discussions with various patent offices’ staff members. Analysis of techniques is based on critical literature review. We also reviewed websites of 11 top IPR tools providers as identified by interviewees and/or the Hyperion MarketView™ Report (2016) and Capterra’s review (2017 ). Analysis is based on reviews of websites (November, 2017) by Anaqua for Corporations, IP One (from CPA Global), InnovationQ (from ip.com), IPfolio, PatentSight, Unycom Enterprise, Wellspring’s IP management software, Patricia (form Patrix), Alt Legal, Inteum, Dennemeyer’s DIAMS iQ [2] .

4. The potential social and IPRM benefits of new advances in the field of IPR

One of the biggest problems of IPR data usability is the rapid growth of number of IPR, especially patents. They are written in different languages and it has become increasingly challenging to understand the state of the art, this consequently causing duplication of research and increasing the number of invalid patents granted. Once errors can be corrected, it will be easier to identify inherently invalid patents previously granted, and consequently leading to a natural rise in the quality of IPR.

Governments have a large quantity of IPR-related data, which can be of economic and social value to society. European Patent Office (EPO) sees the advantages of its new LOD patent databases, one of the outlets of the new open data trend, as increased availability of data from different sources via one channel, less “data friction” when combining different data sets, more effective linking with business information and increased trust thanks to provenance ( Kracker, 2017 ). The Korean Patent Office (KIPO) also saw its efforts in a similar manner ( KIPO, 2016 ).

The growing importance of IPR Open (linked) data is connected to better transparency making it easier for companies to understand their value. However, if we could not only have exploitable open databases, but if these could also be combined with IPR techniques with AI functionality, and additionally, IPR tools which supported the handling of IPR data by integrating some AI functionalities, we could be seeing a new form of tacit knowledge, the “Artificial intelligence knowledge” creation (see Figure 1 ). Therefore, the often problematic issue of tacit knowledge inside the IPR field embodied in individuals (note that the usual way of gaining IPRM, exploitation and other connected IPR knowledge is through apprenticeship and that the rotation of individuals presents a serious problem for especially company IPR departments, Modic and Damij, 2018 )) would be transformed into a latent explicit knowledge (knowledge available on recall as opposed to explicit knowledge, always available). Solutions, like IBM Watson, seem to also be a game changer in this area. Watson identified compounds on which the patent protection has already lapsed, and the pilot results suggest that Watson can accelerate identification of novel drug candidates and novel drug targets by harnessing the potential of patent (and connected) big data ( Chen et al. , 2016 ). The IBM team believes the insights provided by Watson technology are to be used as a guide, i.e., as augmented intelligence – which is capable of ingesting, digesting, understanding and analysing data and can be harnessed in various elements of IPR processes: from evidence of use, to prior art, patent landscapes and portfolio analysis ( Fleischman, 2018 ). If the technology was widely available with all its features, this could present a significant change, as it would enable smaller entities to access knowledge that is now tacit knowledge.

When discussing traceability, blockchain is one of the frequently debated issues. Several potential social benefits, as derived from the utilisation of blockchain in the field of IPR, are present. A tool for registration of IPRs could simplify registration and lower the costs ( Vella et al. , 2018 ; Morabito, 2017 ) or could be an alternative to IPR registration, especially patents. Thus, it has a potential particularly for small entities (independent inventors, SMEs, non-profit organisations), as well as inventors and organisations from less developed countries, who are unable to access the current world patent system simply because it is too expensive for them.

Blockchain provides a robust and trustworthy method of establishing business ownership on intangible assets, including IPR ( Morabito, 2017 ) and thus has the potential to enhance transparency of IPR transactions ( Vella et al. , 2018 ). Not only does this have positive effects for individual companies, but it can also streamline the costs of operations for patent offices, and reduced options for litigation can lower court case numbers and reduce court backlogs. Furthermore, it also has the potential to enable half open licensing, when royalties start only when IPR-based income is generated by downstream users; meaning that without income generation, the half open licenses allow for IPR-based solutions to be spread in an open environment. Moreover, it would allow tracking commons’ knowledge (under open licenses or not) incorporation into corporate IPR portfolios disallowing the privatisation of gains.

With regard to potential IPRM benefits, IPRM deals with managing IPR big data efficiently, and differently ( Braganza et al. , 2017 ; Davenport et al. , 2012 ). McAfee and Brynjolfsson (2012) argue that companies will not reap the full benefits of the transition made in exploiting big data, unless they are able to manage change effectively.

Analysis of the interviews showed a clear trend that IP executives are aware of the growing importance of ICT, and their role in IPRM, however, they continue to struggle with defining how to integrate IPR tools to achieve best outcome. A Senior IP Counsel at a German multinational chemical manufacturing corporation stated that, “IT developments will have a big impact in the near future on IP development, because the more transparent you make the IP, the easier it is for management to understand its value”.

Utilising the ICT in IPR processes is possible, however, doing it in the most efficient way to enable companies to achieve maximum benefits, is the ideal. Some companies use a range of different software tools connected to IPR and IPRM, whilst others try to find or develop software that integrates as many features and data sources as possible and are able to connect to other business processes and databases. Generally, the more comprehensive the tool, the less information is missing, and consequently, the higher the satisfaction level. Nonetheless, some experts, such as the Head of Legal Operations and IP Management at a European multinational pharmaceutical corporation, believe that IPR tools often promise more than they deliver. He states that they, “do not think there are any particularly good IP management tools on the market /…/the whole industry still lacks are real IP management tools, helping to relate to the business value more”. IPR experts are seeking a tool that would, in addition to being a comprehensive docketing system and simple interface retrieval of data from public IPR databases, also encompass supplying or channelling invention disclosures to pertinent individuals, providing functionality for IPR valuation, evaluation and analysis.

The next chapter will provide more detail deal with regard to the technological dimension, providing an analysis on the current state of linked open databases, software tools for IPRM and techniques that support IPR data correction and analytics.

5. Technology

5.1 databases and linked (open) data.

Since the Venetian patent statute of 1474, IPR have retained their connection to the concept of openness and dissemination of ideas in exchange for limited time monopolies. There are various types of databases and online sources connected with IPR constituting Layer 1 in the framework in Table I . Public patent databases as the original sources allow raw data retrieval and the use of interfaces by providing patent texts and some metadata. Related IPR databases include, for example, those related to patent disputes, patent citations. Business databases provide information on IPR owners, etc. Scientific databases provide us inter alia with data on inventors. Miscellaneous online data sources include less or more structured sources, e.g., business news, blogs-based IPR-related texts, information on IPR experts. Multi-source IPR databases provide broader information, e.g., on IPR quality and business connected data. Two examples of the latter are the data set linking the EPO and USPTO patent data to Amadeus business database and the Oxford Firm-Level IP Database ( Thoma and Torrisi, 2007 ; Helmers et al. , 2011 ).

Linked open data (IPR LOD) databases are the latest evolution in IPR databases, although the concept of LOD goes back to 2006, when principles such as using uniform resource identifiers as names for things and including links were put forward ( Berners-Lee, 2006 ). Linked data are data published on the web in a machine-readable format, which can be linked to or from external data ( Bizer et al. , 2009 ). LOD is in essence a format allowing for efficient (multi-source) database utilisation as the term refers to a set of practices for publishing and interlinking structured data ( Auer, 2014 ).

Combining this to ideas of open data, we get LOD, structured data made available for others to be reused ( Mezaour et al. , 2014 ). The concept is connected to the Open Data movement to ensure public government data are accessible in non-proprietary formats ( Bauer and Kaltenböck, 2012 ). However, LOD landscape includes databases provided by non-governmental entities. DBPedia, extracting structured knowledge from Wikipedia, is often seen as the “nucleus” of LOD ( Auer et al. , 2007 ). Furthermore, patent data of individual patent offices are sometimes provided by outside providers, such as in the case of USPTO or (formally) the EPO.

Table I shows the classification of patent offices and their data according to the Berners-Lee Five Star Open Data Plan. More stars indicate data formats more conducive to open data policies, as they allow for easier export and import of data, and more streamlined merging and analysis. The category **** is redundant as there is no standalone RDF providing databases; and, we would suggest an introduction of the *****+ category, where the additional criteria is the existence of linkages with other data, signalling the real uptake of the raw data by users (see Table I ). The Type indicates the most Open data friendly format, though patent offices often provide other formats simultaneously. They often also provide more than one database, and the degree of the export varies for bibliographical data (Swiss Patent Database offering up to 25 variables).

Five patent offices are leading in terms of IPR LOD; USPTO, EPO, KIPO, IPAustralia and IPO UK. Cooperation of national offices with Espacenet was also advantageous, as it produced the option of a limited bibliographic data download in .csv format (not taken into account above). However, most of the patent offices can still be categorised only as Type * or Type **, their data remaining in linkable open data unfriendly formats.

There are only a few databases that could be categorised as *****+, or that have shown other initiatives to make exporting, merging and analysing data easier. For example, KIPO has not only published the IPR LOD, but also included the owners’ corporate registration number and the Australian Patent Office IPR database includes information about companies’ size, technology and geographic location, making it easier for users to link data on patents to information on related business entities ( KIPO, 2016 ; Man, 2014 ).

Currently, EPO’s Linked open data is the newest of the few IPR LOD databases at users’ disposal. It builds upon their previous work in connecting patent-related data, such as their Deep Linking service, allowing users to consult the EP document’s legal status data. However, the IPR LOD database remains as a raw data product and without additional skills and resources cannot be fully utilised, which could potentially widen the gap between SMEs and IPR-savvy companies. For example, the linkage to DBPedia has also been carried out, but since then de-installed ( Kracker, 2017 ). This year the EPO also included in their Research grant call explicitly the field of linked open data and solutions therein, where at least one project will start end of this year linking EPO database with the Springer database ( IP LodB, 2018 ). The current LOD IPR landscape shown below is based on the The Linking Open Data cloud diagram and upgraded [3] .

Figure 2 shows patent LOD databases [4] we could call *****+, and their inbound and outbound links, as per The Linking Open Data cloud diagram ( LOD cloud, 2018 ) – a complex LOD ecosystem currently listing 1,164 data sets. They are also linked to the most inbound and outbound link-rich LOD databases, namely, the Comprehensive KAN and DBPedia. The new EP LOD and KIPO databases have no data on linkages, even though some attempts were made as mentioned above. There are, however, several LOD databases that this patent data could be linked to; e.g. the recently published bibliographic LOD database by Springer Nature SciGraph or the older New York Times LOD.

When considering the traceability of IPR data, some patent offices offer centralised solutions, such as i-DEPOT, which allows to trace the date of inventions’ creation. However, at the forefront of these debates is blockchain as a disruptive technology, due to its transparency, decentralisation and prevention of infringements and fraud. Blockchain is a chain of blocks of chronologically linked information, replicated in a distributed database. Information can be added, but never removed, changes are registered and validated. Individual blocks can be protected by cryptography, and only those authorised can access the information ( McPhee and Ljutic, 2017 ). Blockchain application to IPR can be either inside the registration or exploitation phases (related to issues of licensing, proving authenticity and piracy) ( Vella et al. , 2018 ; Morabito, 2017 ) as well as distribution. In case of licensing, the topic is connected to smart contracts, open licenses and IPR-based collaboration ( Pilkington, 2016 ; Morabito, 2017 ). Smart contracts are computer codes that reside in the blockchain and are implemented if certain conditions are met, which is confirmable by a number of computers to ensure truthfulness ( Morabito, 2017 ; Szabo, 1997 ). There are numerous potential applications of blockchain connected to IPR. Also, the Linked Data paradigm is evolving from an academic concept for addressing one of the biggest challenges in the area of information management the exploitation of the web as a platform for data and information integration; to practical applications in IPR field deriving from the transfer from the Web of Documents to a Web of Data. Yet, it is clear there is still much to be done, both in terms of the volume of IPR LOD-connected databases, as well as their functionality in linking to other LOD data sets as well as the real-life uptake of blockchain solutions.

5.2 Classification of tools and techniques

This chapter summarises the techniques and tools (technology Layers 2 and 3 as set out in Figure 1 ) that analyse large quantities of patent documents and other IPR data to provide useful information to various users.

The EPO’s database, Espacenet, on its own, currently contains over 100m patent documents from 90 patent authorities worldwide. Whilst patent data are exceptionally important, it is also very difficult to extract some useful information from it as patents are mostly stored as images; written in different languages; countries have different patent requirements; no uniform structural requirements; some patent figures are drawn by hand, some on computer; some patent attorneys intentionally use misleading language; incomprehensible language and grammatical mistakes can be also used inadvertently. How to deal with these issues remains a challenge.

There are several possible taxonomies of IPR software. Considering their functionalities we see tools supporting different phases of the innovation cycle, those supporting financial management (record and estimate costs), archiving documents (IPR portfolio) and enabling communication between users and IPR offices. Some tools have functionality to integrate data from external databases, such as patent litigation information and patent citation indexes. In terms of intended user-base we have IPR tools for companies, for IPR experts and for technology transfer offices.

There is an upward trend in the creation of new IPRM software in recent years. However, after reviewing the websites of the 13 most important IPR tools providers by Hyperion MarketView™ Report (2016) it appears that these tools only modestly respond to the challenges raised, and largely look like any project management software. Bonino et al. (2010) was optimistic with regard to semantic-based solutions, however, some of the tools he describes are currently in poor condition or unavailable.

In terms of techniques utilised in semantic analysis, Abbas et al. (2014) made a taxonomy of proposed computer-assisted patent analysis techniques where they distinguish between text mining and visualisation approaches. These two categories are based on frequent use-cases, whilst the underlying methods are primarily inspired by IR and ML. This is not unreasonable, as patent documents are similar to other types of documents in that they contain textual and visual data as well as references to other documents.

As seen in Figure 3 , a typical IR system consists of document pre-processing, feature extraction and feature analysis. Each of those steps can be based on heuristic rules or utilise machine learning methods. In the following paragraphs, we review the use of different techniques in the IPR research domain in the last decade, with a particular focus on the works referenced in recent literature reviews by Abbas et al. (2014) and Aristodemou and Tietze (2017) . The list is by no means complete, it is only focussed on key examples illustrating the diversity and potential of such methods.

The patent document pre-processing step involves scanning the unstructured data (text and images) and extracting useful information from it.

Due to the nature of the patent data, the approaches mainly focus around text mining techniques; meaning using some kind of natural language processing ( Wang et al. , 2015 ; Han et al. , 2017 ), such as subject–action–object analysis ( Park, Kim, Choi and Yoon, 2013 ; Park, Ree and Kim, 2013 ), property–function analysis ( Dewulf, 2013 ) or rule-based analysis to extract semantic primitives. Several authors have also proposed the utilisation of patent images and sketches in patent analysis, in order to determine similarities between patents ( Bhatti and Handbury, 2013 ). In terms of pre-processing, image analysis challenges involve localisation of images and sub-images, categorisation of images and label recognition ( Vrochidis et al. , 2010 ). The primary sources of inter-information are cross-patent citations ( Altuntas et al. , 2015 ).

The feature extraction methods transform low-level semantic primitives into a document-wide representation. By involving projection of each document into a high-dimensional feature space we can determine bounds between classes or proximity of documents. When processing textual data, the semantic primitives can be frequency vectors ( Chen and Yu-Ting, 2011 ), vectors of concepts that describe higher-level semantic information, or domain-specific hierarchical structures ( Lee, 2013 ). In analysis of patent sketches, content is frequently encoded with shape or texture descriptors ( Bhatti and Handbury, 2013 ) due to the line-art nature of visual information.

The method used in the feature analysis stage depends on the problem at hand, for example, retrieval of similar patents. In this case, IR techniques based on vector distances ( Lee, 2013 ) are used to infer which documents are most similar. Another task is automatic classification of patents using ML methods. Scenarios include patent quality analysis ( Wu et al. , 2016 ), patent categorisation ( Vrochidis et al. , 2010 ) and determining the impact of patents on other aspect of companies ( Chen et al. , 2013 ). Supervised learning methods, such as support vector machines ( Wu et al. , 2016 ) or artificial neural networks ( Chen et al. , 2013 ), are frequently used in such cases. In explorative analysis of the patent landscape for trend identification, people have also utilised unsupervised learning methods, like clustering ( Atzmüller and Landl, 2009 ; Madani and Weber, 2016 ) and network analysis ( Dotsika, 2017 ; Park, Kim, Choi and Yoon, 2013 ).

Despite the apparent contribution of IR methods in transforming access to information, they are harder to apply to semantic-sensitive fields, such as IPR analysis, with the same level of success. The crucial information in patent documents can be difficult to extract automatically because of objective (history, language) or subjective (intentional misuse of description) reasons. As noted by Lupu (2017) , the level of research interest in this field has, after more than a decade of increasing optimism, decreased in the past years. This can be in part attributed to the realisation that extracting high-level semantic content from sophisticated unstructured text and images is very a challenging problem. The most successful working cognitive computing system is IBM Watson, who has already been analysing patent information in the past, with a particular emphasis in the pharmaceutical sector. However, this system is proprietary and accessible only to a limited number of influential clients.

6. Discussion

Over the last years, activities around IPR Open Data, merging of IPR data with related data, IPR Linked Data, IPR-linked open databases and the debates over utilising the Semantic Web opportunities have gained momentum. However, this should go hand in hand with organisations (both public and private) publishing structured data (complying also with linked data standards/principles), the advances in new techniques, as well as IPR tools and their increased availability. Companies and other patent and IPR data users need to draw on those advanced technologies and tools in order to combine, query (and analyse) data as part of their business intelligence, as well as to improve their services and products.

In terms of the availability of data, the amount of IPR and IPR-connected data publically available is increasing. Responding to P1 , the new trends towards formats supporting more export-ready, merge-ready and analysis-ready data are also real, although the amount of patent data available (e.g. as LOD) is still relatively low. LOD means the data are “linkable”, not that it is already linked. This means that the uptake of these databases by the users can be slow and can even widen the gap between the IPR-savvy multinationals with sufficient resources and other smaller entities and individuals. The latter would defeat the purpose of publishing such databases, if the objective was to make IPR data more useful to more groups of users, especially also non-patent savvy users (data scientists, web developers, companies integrating IP into their products). Some steps are taken towards this, for example, IPNOVA (available at the moment as a beta version) which is the interface to the IPAustralia’s IPGOD database. Another route (contrasting somewhat with developing interfaces) is through sufficient dissemination and training workshop accompanying the releases of databases in new formats. On the other hand, the authors remain hopeful as new entities – including private and NGO entities – provide more and more LOD databases, and with growth of potential links, allowing greater potential for IPR.

In response to P2 , techniques that would support IPR data correction, and IPR data analytics and software tools, which support IPRM, are still not at a sufficient stage of development for IPR managers and other users dealing with IPR. The IPR tools remain primarily visualisation tools ( P3 ); or project management and docketing tools, applied to the field of IPR. There are few true IPRM tools that also integrate variable (external and internal) data merges and harness new advances in IPR techniques, although some solutions have been integrated. This is perhaps because the existing techniques, which are suitable for many existing retrieval and analysis tasks, are frequently branded as “AI”, a term that increases expectations about the capabilities which existing methods fail to fulfil. A complete AI system is perhaps the ultimate goal of automatic patent analysis, capable of high-level reasoning about the content of patent documents, comparing their underlying ideas and determining similarities. The current state is (far) removed from this goal. At present, it is primarily addressing very narrow domains, interpretable by data scientists and machine learning researchers. However, as also noted by Lupu (2017) , recent breakthroughs in deep learning and artificial neural networks already address tasks such as machine translation and image analysis, which can be (and sometimes are) utilised in IPR analysis.

In response to P4 , blockchain technology is now fairly widely discussed for its potential to change the nature of IPRs by simplifying registration, lowering costs, increasing transparency and enabling or improving licensing and other transfers of IPR. However, the technology has certain limitations and still needs significant time to develop. This is not only because of the influence that transnational companies have on policy makers, but also, the technology itself might have some weaknesses. It needs huge processing power and therefore for now requires high-volume electricity consumption. Second, field, such as the IPR field, has its own set of limitations connected to legal and judicial frameworks. Therefore, it is important to carefully determine fields where it would be used. “Despite the many interesting potential uses of blockchain technology, one of the most important skills in the developing industry is to see where it is and is not appropriate to use cryptocurrency and blockchain models” ( Swan, 2015 ). Although there are various social and IPRM benefits of employing blockchain technologies in the field of IPR, caution must be applied.

To conclude, despite significant efforts in the last decades, in the field of information technology support to IPRs, and the more and more used buzzwords of augmented intelligence and augmented expertise also for IPR, there is more time needed before these progressive ideas will become (widespread) reality.

Technology and benefits in IPR landscapes

Narrow IPR LOD landscape (patent databases)

A typical computer-assisted document analysis pipeline as IPR techniques classification framework

Classification of patent offices according to the Berners-Lee Open Data Plan

Notes: a Taking into account the AKSW database (different provider); b the patent offices have done additional steps non-related to the format to make merging of data easier; c the database can be described as providing linking data, yet it is not an LOD database in classical sense; d if taking into account the bibliographical export in .csv by Espacenet on its web-pages designed in cooperation with national patent offices (e.g. https://sk.espacenet.com/ ), there are such data provided for most, however, the end document exports remain .pdf

Missing from the list are the Latvian, Icelandic, Maltese and Cyprus Patent Office, as they only refer to Espacenet or there is a lack of sufficient information. The classification takes into account data that is (formally) provided by outside sources (e.g. for USPTO).

We have also taken into account a review of the available semantic solutions that was made by Bonino et al. (2010 , p. 37, Table 9). However, these new technology enablers are currently in a less than ideal state (in poor condition or unavailable) and they (those which are at least available) look more like a scientific experiment than a final product that would support real patent analytics in companies. Though we sent some follow-up e-mails we did not receive much useful information so they were excluded from the paper.

Eito-Brun (2015) lists 31 LOD databases according to datahub.io related to patents, but they could be hardly classified as IPR databases.

The Linked Open Data Cloud diagram includes EPO reference, which was created and published by the research group AKSW.

Abbas , A. , Zhang , L. and Khan , U.S. ( 2014 ), “ A literature review on the state-of-the-art in patent analysis ”, World Patent Information , No. 37 , pp. 3 - 13 .

Altuntas , S. , Dereli , T. and Kusiak , A. ( 2015 ), “ Analysis of patent documents with weighted association rules ”, Technological Forecasting and Social Change , No. 92 , pp. 249 - 262 .

Aristodemou , L. and Tietze , F. ( 2017 ), “ A literature review on the state-of-the-art on intellectual property analytics ”, Working Paper Series No. 2, Centre for Technology Management, Vol. 2017, Cambridge, pp. 1-13, available at: www.repository.cam.ac.uk/handle/1810/268007 (accessed 25 August 2018 ).

Atzmüller , P. and Landl , G. ( 2009 ), “ Semantic enrichment and added metadata – examples of efficient usage in an industrial environment ”, World Patent Information , Vol. 31 No. 2 , pp. 89 - 96 .

Auer , S. ( 2014 ), “ Introduction to LOD2 ”, in Auer , S. , Bryl , V. and Tramp , S. (Eds), Linked Open Data – Creating Knowledge Out of Interlinked Data Results of the LOD2 Project , Springer , Cham, Heidelberg, New York, NY, Dordrecht and London , pp. 1 - 20 .

Auer , S. , Bizer , C. , Kobilarov , G. , Lehmann , J. , Cyganiak , R. and Ives , Z. ( 2007 ), “ DBpedia: a nucleus for a web of open data ”, in Aberer , K. et al. (Eds), ISWC/ASWC 2007 , Springer-Verlag , Berlin and Heidelberg , pp. 722 - 735 .

Bauer , F. and Kaltenböck , M. ( 2012 ), Linked Open Data: The Essentials – A Quick Start Guide for Decision Makers , Edition mono/monochrom , Vienna .

Berners-Lee , T. ( 2006 ), “ Linked data: a personal view note ”, available at: www.w3.org/DesignIssues/LinkedData.html (accessed 30 January 2018 ).

Bhatti , N. and Hanbury , A. ( 2013 ), “ Image search in patents: a review ”, International Journal on Document Analysis and Recognition , Vol. 16 No. 4 , pp. 309 - 329 .

Bizer , C. , Heath , T. and Berners-Lee , T. ( 2009 ), “ Linked data – the story so far ”, International Journal on Semantic Web and Information Systems , Vol. 5 No. 3 , pp. 1 - 22 .

Bonino , D. , Ciaramella , A. and Corno , F. ( 2010 ), “ Review of the state-of-the-art in patent information and forthcoming evolutions in intelligent patent informatics ”, World Patent Information , No. 32 , pp. 30 - 38 .

Braganza , A. , Brooks , L. , Nepelski , D. , Ali , M. and Moro , R. ( 2017 ), “ Resource management in big data initiatives: processes and dynamic capabilities ”, Journal of Business Research , No. 70 , pp. 328 - 337 .

Capterra ( 2017 ), “ Captera review ”, available at: www.capterra.com/ (accessed 10 December 2017 ).

Chen , Y. and Yu-Ting , C. ( 2011 ), “ An IPC-based vector space model for patent retrieval ”, Information Processing and Management , Vol. 47 No. 3 , pp. 309 - 322 .

Chen , Y. , Argentinis , E. and Weber , G. ( 2016 ), “ IBM Watson: how cognitive computing can be applied to big data challenges in life sciences research ”, Clinical Therapeutics , Vol. 38 No. 4 , pp. 688 - 701 .

Chen , Y.S. , Tien , W.P. , Chen , Y.W. , Lin , C.C. and Lee , Y.I. ( 2013 ), “ Using artificial neural network (ANN) to explore the influences of number of inventors, average age of patents, and age of patenting activities on patent performance and corporate performance ”, Proceedings – 2013 4th World Congress on Software Engineering , pp. 136 - 139 , available at: https://ieeexplore.ieee.org/document/6754276 (accessed 25 August 2018 ).

Ciccatelli , A. ( 2017 ), “ The future of big data and intellectual property ”, available at: www.insidecounsel.com/2017/06/13/the-future-of-Big-Data-and-intellectual-property?&slreturn=1507773271 (accessed 15 May 2018 ).

Davenport , T.H. , Barth , P. and Bean , R. ( 2012 ), “ How ‘Big Data’ is different ”, MIT Sloan Management Review , Vol. 54 No. 1 , pp. 43 - 46 .

Dewulf , K. ( 2013 ), “ Sustainable product innovation: the importance of the front end stage in the innovation process ”, in Coelho , D. (Ed), Advances in Industrial Design Engineering , Intech Open , Rijeka , pp. 139 - 166 .

Dotsika , F. ( 2017 ), “ Identifying potentially disruptive trends by means of keyword network analysis ”, Technological Forecasting and Social Change , No. 119 , pp. 114 - 127 .

Eito-Brun , R. ( 2015 ), “ Patents, semantics and open innovation: the role of LOD in a business directory for knowledge intensive industries ”, Presentation at Dr Haxel Congress and Event Management GmbH, Nice, 20 October, available at: www.slideshare.net/Haxel/linked-open-data-in-the-world-of-patents (accessed 10 February 2018 ).

European Data Portal ( 2017 ), Open Data Maturity in Europe 2017: Open Data for a European Data Economy , Publications Office of the European Union , Luxembourg , available at: www.europeandataportal.eu/sites/default/files/edp_creating_value_through_open_data_0.pdf (accessed 14 February 2018 ).

Fleischman , T. ( 2018 ), “ Augmented intelligence, intellectual property and IBM ”, Presentation at AI: Hype vs Reality and the Impact on the Patent Industry Clarivate Analytics Online Webinar .

Freeman , C. and Louca , F. ( 2002 ), As Time Goes By: From the Industrial Revolution to the Information Revolution , Oxford University Press , Oxford .

Han , Q. , Heimerl , F. , Codina-Filba , J. , Lohmann , S. , Wanner , L. and Ertl , T. ( 2017 ), “ Visual patent trend analysis for informed decision making in technology management ”, World Patent Information , No. 49 , pp. 34 - 42 .

Helmers , C. , Rogers , M. and Schautschick , P. ( 2011 ), “ Intellectual property at the firm-level in the UK: the Oxford firm-level intellectual property database ”, Discussion Paper Series No. 546, Department of Economics, University of Oxford, Oxford, pp. 1-23, available at: www.economics.ox.ac.uk/materials/working_papers/paper546.pdf (accessed 14 February 2018 ).

Hyperion MarketView™ Report ( 2016 ), “ Anaqua enterprise as featured in the Hyperion MarketView™ research program intellectual property management systems for corporation (2016) ”, available at: www.anaqua.com/anaqua-hyperion-report.html (accessed 25 February 2018 ).

IP LodB ( 2018 ), “ Granted project proposal IP linked open data: building bridges (IP LodB) ”, IP LodB (forthcoming), Muenich , pp. 1 - 96 .

KIPO ( 2016 ), “ Korean Intellectual Property Office and Patent Information Expansion of providing public services (English title) ”, available at: www.kipo.go.kr/kpo/user.tdf?a=user.news.press1.BoardApp&board_id=press&c=1003&catmenu=m03_05_01&seq=15589 (accessed 13 February 2018 ).

Kracker , M. ( 2017 ), “ New PI product: linked open EP data ”, Presentation at IP Statistics for Decision Makers Conference , Mexico City , 14–15 November .

Lee , S. ( 2013 ), “ How to assess patent infringement risks: a semantic patent claim analysis using dependency relationships ”, Journal Technology Analysis & Strategic Management , Vol. 25 No. 1 , pp. 23 - 38 .

LOD cloud ( 2018 ), “ The linked open data cloud ”, available at: https://lod-cloud.net/ (accessed 20 December 2018 ).

Lupu , M. ( 2017 ), “ Information retrieval, machine learning, and natural language processing for intellectual property information ”, World Patent Information , Vol. 49 , pp. A1 - A3 .

McAfee , A. and Brynolfsson , E. ( 2012 ), “ Big data: the management revolution, ” Harvard Business Review , Vol. 90 No. 10 , pp. 60 - 68 .

McPhee , C. and Ljutic , A. ( 2017 ), “ Editorial: blockchain ”, Technology Innovation Management Review , Vol. 7 No. 10 , pp. 3 - 5 .

Madani , F. and Weber , C. ( 2016 ), “ The evolution of patent mining: applying bibliometrics analysis and keyword network analysis ”, World Patent Information , No. 46 , pp. 32 - 48 .

Man , B. ( 2014 ), “ Overview of the intellectual property government open data ”, IP Australia Economic Research Paper No. 2, pp. 1-18, available at: www.ipaustralia.gov.au/about-us/reports/economics_research_paper02 (accessed 14 February 2018 ).

Mezaour , A. , van Nuffelen , B. and Blaschke , C. ( 2014 ), “ Building enterprise ready applications using linked open data ”, in Auer , S. , Bryl , V. and Tramp , S. (Eds), Linked Open Data – Creating Knowledge Out of Interlinked Data Results of the LOD2 Project , Springer , Cham, Heidelberg, New York, NY, Dordrecht and London , pp. 155 - 174 .

Modic , D. ( 2017 ), “ Intellectual property rights in the information sector: sociological perspective ”, in Rončević , B. and Tomšić , M. (Eds), Information Society and Its Manifestations: Economy, Politics, Culture , Peter Lang , Frankfurt am Main , pp. 55 - 70 .

Modic , D. and Damij , N. ( 2018 ), Towards Intellectual Property Management: Back Office and Front Office Perspectives , Palgrave Macmillan , Cham .

Morabito , V. ( 2017 ), Business Innovation through Blockchain: The B 3 Perspective , Springer Nature , Cham .

Park , H. , Ree , J.J. and Kim , K. ( 2013 ), “ Identification of promising patents for technology transfers using TRIZ evolution trends ”, Expert Systems with Applications , Vol. 40 No. 2 , pp. 736 - 743 .

Park , H. , Kim , K. , Choi , S. and Yoon , J. ( 2013 ), “ A patent intelligence system for strategic technology planning ”, Expert Systems with Applications , Vol. 40 No. 7 , pp. 2373 - 2390 .

Pilkington , M. ( 2016 ), “ Blockchain technology: principles and applications ”, in Olleros , X.F. and Zhengu , M. (Eds), Research Handbook on Digital Transformations , Edward Elgar Publishing , Cheltenham and Northampton , pp. 225 - 253 .

Stading , T. ( 2017 ), “ Using big data to make intellectual property a strategic weapon ”, available at: https://dzone.com/articles/using-Big-Data-to-make-intellectual-property-a-str (accessed 10 February, 2018 ).

Swan , M. ( 2015 ), Blockchain – Blueprint for a New Economy , O’Reilly Media , Sebastopol, CA .

Szabo , N. ( 1997 ), “ Formalizing and securing relationships on public networks ”, First Monday , Vol. 2 No. 9 , pp. 1 - 8 , available at: http://firstmonday.org/article/view/548/469 (accessed 29 January 2018 ).

Thoma , G. and Torrisi , S. ( 2007 ), “ Creating powerful indicators for innovation studies with approximate matching algorithms: a test based on PATSTAT and Amadeus databases ”, Working Paper No. 211, CESPRI Bocconi University, Milan, available at: https://ideas.repec.org/p/cri/cespri/wp211.html (accessed 14 February 2019 ).

Vella , D. , Falzon , M. , Cassar , T. and Valenzia , A. ( 2018 ), “ Blockchain’s applicability to intellectual property management ”, The Licensing Journal , Vol. 38 No. 1 , pp. 10 - 12 .

Vrochidis , S. , Papadopoulos , S. , Moumtzidou , A. , Sidiropoulos , P. , Pianta , E. and Kompatsiaris , I. ( 2010 ), “ Towards content-based patent image retrieval: a framework perspective ”, World Patent Information , Vol. 32 No. 2 , pp. 94 - 106 .

Wang , J. , Lu , W.F. and Loh , H.T. ( 2015 ), “ A two-level parser for patent claim parsing ”, Advanced Engineering Informatics , Vol. 29 No. 3 , pp. 431 - 439 .

White , E. ( 2018 ), “ AI in patent research ”, Presentation at AI: Hype vs Reality and the Impact on the Patent Industry Clarivate Analytics Online Webinar .

Wu , J.L. , Chang , P.C. , Tsao , C.C. and Fand , C.Y. ( 2016 ), “ A patent quality analysis and classification system using self-organizing maps with support vector machine ”, Applied Soft Computing , No. 41 , pp. 305 - 316 .

Further reading

Lee , S. , Yoon , B. , Lee , C. and Park , J. ( 2009 ), “ Business planning based on technological capabilities: patent analysis for technology-driven roadmapping ”, Technological Forecasting and Social Change , Vol. 76 No. 6 , pp. 769 - 786 .

Acknowledgements

Dr Damij would like to acknowledge the ARRS Grant No. ARRS-P1-0383(A). Dr Hafner would like to acknowledge Operation No. C3330-17-529006 “Researchers-2.0-FIŠ-529006” supported by ERDF and Republic of Slovenia, Ministry of Education, Science and Sport. Dr Modic would like to acknowledge the JSPS International Research Grant ID No. 16774 and JSPS KAKENHI Grant No. 16F16774. Dr Hafner and Dr Modic acknowledge that this paper has been co-funded by the Academic Research Programme of the European Patent Office. The research results contained in this paper are those of the researchers only. They do not necessarily represent the views of the EPO.

Corresponding author

Related articles, we’re listening — tell us what you think, something didn’t work….

Report bugs here

All feedback is valuable

Please share your general feedback

Join us on our journey

Platform update page.

Visit emeraldpublishing.com/platformupdate to discover the latest news and updates

Questions & More Information

Answers to the most commonly asked questions here

• AI and IP Policy
• AI and IP Strategy Clearing House
• Virtual Experience

Artificial Intelligence and Intellectual Property

Artificial intelligence (Al) is increasingly driving important developments in technology and business. It is being employed across a wide range of industries with impact on almost every aspect of the creation. The availability of large amounts of training data and advances in affordable high computing power are fueling Al's growth. Al intersects with intellectual property (IP) in a number of ways.

There is no universal definition of artificial intelligence (AI). AI is generally considered to be a discipline of computer science that is aimed at developing machines and systems that can carry out tasks considered to require human intelligence. Machine learning and deep learning are two subsets of AI. In recent years, with the development of new neural networks techniques and hardware, AI is usually perceived as a synonym for “deep supervised machine learning”.

Machine learning uses examples of input and expected output (so called “structured data” or “training data”), in order to continually improve and make decisions without being programmed how to do so in a step-by-step sequence of instructions. This approach mimics actual biological cognition: a child learns to recognize objects (such as cups) from examples of the same objects (such as various kinds of cups). Today application of machine learning are widespread including email spam filtering, machine translation, voice, text and image recognition.

WIPO has implemented and continues to develop its own IP management services and tools using AI technologies, creating best-in-class applications of AI for the international IP system. Find out more about WIPO AI tools .

Read the full list of AI FAQs .

Summary of the Sixth session is available .

Sixth Session of the WIPO Conversation “Frontier technologies – AI Inventions”

AI inventions present the current patent system with a number of challenges. In order to provide better protection and embrace the full value of patents it is necessary to have timely, transparent and accessible standards for patent granting that market players can fully rely on. To harness the economic potential of AI it is important to understand the uncertainties faced by innovators and for IP Offices to consider how to best support AI innovation.

What are the market trends and how do these translate in terms of patent applications? How autonomous is AI actually, what role does it play either as part of the inventive process or as part of an invention? What questions does this raise for the IP system? And how are IP Offices supporting AI inventors?

The Sixth Session of the WIPO Conversation took place on September 21-22, 2022 and connected the views and stories from innovators and IP professionals to what is happening in the IP Offices around the world and aimed at sharing information and building awareness around patent examination practices, tools and guidelines for AI inventions.

First three sessions of the WIPO Conversation

The  first three sessions of the WIPO Conversation looked on the broad conceptual questions raised by AI for the IP system, with a view to collectively formulating the questions that policymakers need to ask.

AI and IP Clearing House

AI is becoming a strategic capability for many governments across the globe. Strategies for the development of AI capacity and AI regulatory measures are being adopted with increasing frequency.

WIPO continuously collates and publishes the main government instruments of relevance to AI and IP with the aid of the Member States. Member States are invited to inform WIPO about any updates in their policies.

Open discussion and dialogue with Member States on IP and Frontier Technologies and AI

What’s next? WIPO General Assemblies’ side event!

Find out more

WIPO, AI and IP: A Virtual Experience (archived)

The online exhibition, "WIPO: AI and IP, A Virtual Experience" held from September 2020 to March 2021 sought to foster a wider understanding of the relationship between IP Policy and AI and focused on examples of the use of AI in art, music, technology and entertainment.

Watch the archived version of the Exhibition

Introduction

Entertainment

• Browse All Articles
• Newsletter Sign-Up

IntellectualProperty →

No results found in working knowledge.

• Were any results found in one of the other content buckets on the left?
• Try removing some search filters.
• Use different search filters.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• Advanced Search
• Journal List
• J Adv Pharm Technol Res
• v.2(2); Apr-Jun 2011

Intellectual property rights: An overview and implications in pharmaceutical industry

Chandra nath saha.

Quality Assurance Department, Claris Lifesciences Ltd., Ahmedabad, Gujarat, India

Sanjib Bhattacharya

1 Pharmacognosy Division, Bengal School of Technology (A College of Pharmacy), Sugandha, Hooghly, West Bengal, India

Intellectual property rights (IPR) have been defined as ideas, inventions, and creative expressions based on which there is a public willingness to bestow the status of property. IPR provide certain exclusive rights to the inventors or creators of that property, in order to enable them to reap commercial benefits from their creative efforts or reputation. There are several types of intellectual property protection like patent, copyright, trademark, etc. Patent is a recognition for an invention, which satisfies the criteria of global novelty, non-obviousness, and industrial application. IPR is prerequisite for better identification, planning, commercialization, rendering, and thereby protection of invention or creativity. Each industry should evolve its own IPR policies, management style, strategies, and so on depending on its area of specialty. Pharmaceutical industry currently has an evolving IPR strategy requiring a better focus and approach in the coming era.

INTRODUCTION

Intellectual property (IP) pertains to any original creation of the human intellect such as artistic, literary, technical, or scientific creation. Intellectual property rights (IPR) refers to the legal rights given to the inventor or creator to protect his invention or creation for a certain period of time.[ 1 ] These legal rights confer an exclusive right to the inventor/creator or his assignee to fully utilize his invention/creation for a given period of time. It is very well settled that IP play a vital role in the modern economy. It has also been conclusively established that the intellectual labor associated with the innovation should be given due importance so that public good emanates from it. There has been a quantum jump in research and development (R&D) costs with an associated jump in investments required for putting a new technology in the market place.[ 2 ] The stakes of the developers of technology have become very high, and hence, the need to protect the knowledge from unlawful use has become expedient, at least for a period, that would ensure recovery of the R&D and other associated costs and adequate profits for continuous investments in R&D.[ 3 ] IPR is a strong tool, to protect investments, time, money, effort invested by the inventor/creator of an IP, since it grants the inventor/creator an exclusive right for a certain period of time for use of his invention/creation. Thus IPR, in this way aids the economic development of a country by promoting healthy competition and encouraging industrial development and economic growth. Present review furnishes a brief overview of IPR with special emphasis on pharmaceuticals.

BRIEF HISTORY

The laws and administrative procedures relating to IPR have their roots in Europe. The trend of granting patents started in the fourteenth century. In comparison to other European countries, in some matters England was technologically advanced and used to attract artisans from elsewhere, on special terms. The first known copyrights appeared in Italy. Venice can be considered the cradle of IP system as most legal thinking in this area was done here; laws and systems were made here for the first time in the world, and other countries followed in due course.[ 4 ] Patent act in India is more than 150 years old. The inaugural one is the 1856 Act, which is based on the British patent system and it has provided the patent term of 14 years followed by numerous acts and amendments.[ 1 ]

Types of Intellectual Properties and their Description

Originally, only patent, trademarks, and industrial designs were protected as ‘Industrial Property’, but now the term ‘Intellectual Property’ has a much wider meaning. IPR enhances technology advancement in the following ways:[ 1 – 4 ]

• (a) it provides a mechanism of handling infringement, piracy, and unauthorized use
• (b) it provides a pool of information to the general public since all forms of IP are published except in case of trade secrets.

IP protection can be sought for a variety of intellectual efforts including

• (i) Patents
• (ii) Industrial designs relates to features of any shape, configuration, surface pattern, composition of lines and colors applied to an article whether 2-D, e.g., textile, or 3-D, e.g., toothbrush[ 5 ]
• (iii) Trademarks relate to any mark, name, or logo under which trade is conducted for any product or service and by which the manufacturer or the service provider is identified. Trademarks can be bought, sold, and licensed. Trademark has no existence apart from the goodwill of the product or service it symbolizes[ 6 ]
• (iv) Copyright relates to expression of ideas in material form and includes literary, musical, dramatic, artistic, cinematography work, audio tapes, and computer software[ 7 ]
• (v) Geographical indications are indications, which identify as good as originating in the territory of a country or a region or locality in that territory where a given quality, reputation, or other characteristic of the goods is essentially attributable to its geographical origin[ 8 ]

A patent is awarded for an invention, which satisfies the criteria of global novelty, non-obviousness, and industrial or commercial application. Patents can be granted for products and processes. As per the Indian Patent Act 1970, the term of a patent was 14 years from the date of filing except for processes for preparing drugs and food items for which the term was 7 years from the date of the filing or 5 years from the date of the patent, whichever is earlier. No product patents were granted for drugs and food items.[ 9 ] A copyright generated in a member country of the Berne Convention is automatically protected in all the member countries, without any need for registration. India is a signatory to the Berne Convention and has a very good copyright legislation comparable to that of any country. However, the copyright will not be automatically available in countries that are not the members of the Berne Convention. Therefore, copyright may not be considered a territorial right in the strict sense. Like any other property IPR can be transferred, sold, or gifted.[ 7 ]

Role of Undisclosed Information in Intellectual Property

Protection of undisclosed information is least known to players of IPR and also least talked about, although it is perhaps the most important form of protection for industries, R&D institutions and other agencies dealing with IPR. Undisclosed information, generally known as trade secret or confidential information, includes formula, pattern, compilation, programme, device, method, technique, or process. Protection of undisclosed information or trade secret is not really new to humanity; at every stage of development people have evolved methods to keep important information secret, commonly by restricting the knowledge to their family members. Laws relating to all forms of IPR are at different stages of implementation in India, but there is no separate and exclusive law for protecting undisclosed information/trade secret or confidential information.[ 10 ]

Pressures of globalisation or internationalisation were not intense during 1950s to 1980s, and many countries, including India, were able to manage without practising a strong system of IPR. Globalization driven by chemical, pharmaceutical, electronic, and IT industries has resulted into large investment in R&D. This process is characterized by shortening of product cycle, time and high risk of reverse engineering by competitors. Industries came to realize that trade secrets were not adequate to guard a technology. It was difficult to reap the benefits of innovations unless uniform laws and rules of patents, trademarks, copyright, etc. existed. That is how IPR became an important constituent of the World Trade Organization (WTO).[ 11 ]

Rationale of Patent

Patent is recognition to the form of IP manifested in invention. Patents are granted for patentable inventions, which satisfy the requirements of novelty and utility under the stringent examination and opposition procedures prescribed in the Indian Patents Act, 1970, but there is not even a prima-facie presumption as to the validity of the patent granted.[ 9 ]

Most countries have established national regimes to provide protection to the IPR within its jurisdiction. Except in the case of copyrights, the protection granted to the inventor/creator in a country (such as India) or a region (such as European Union) is restricted to that territory where protection is sought and is not valid in other countries or regions.[ 1 ] For example, a patent granted in India is valid only for India and not in the USA. The basic reason for patenting an invention is to make money through exclusivity, i.e., the inventor or his assignee would have a monopoly if,

• (a) the inventor has made an important invention after taking into account the modifications that the customer, and
• (b) if the patent agent has described and claimed the invention correctly in the patent specification drafted, then the resultant patent would give the patent owner an exclusive market.

The patentee can exercise his exclusivity either by marketing the patented invention himself or by licensing it to a third party.

The following would not qualify as patents:

• (i) An invention, which is frivolous or which claims anything obvious or contrary to the well established natural law. An invention, the primary or intended use of which would be contrary to law or morality or injurious to public health
• (ii) A discovery, scientific theory, or mathematical method
• (iii) A mere discovery of any new property or new use for a known substance or of the mere use of a known process, machine, or apparatus unless such known process results in a new product or employs at least one new reactant
• (iv) A substance obtained by a mere admixture resulting only in the aggregation of the properties of the components thereof or a process for producing such substance
• (v) A mere arrangement or re-arrangement or duplication of a known device each functioning independently of one another in its own way
• (vi) A method of agriculture or horticulture
• (vii) Any process for the medicinal, surgical, curative, prophylactic diagnostic, therapeutic or other treatment of human beings or any process for a similar treatment of animals to render them free of disease or to increase their economic value or that of their products
• (viii) An invention relating to atomic energy
• (ix) An invention, which is in effect, is traditional knowledge

Rationale of License

A license is a contract by which the licensor authorizes the licensee to perform certain activities, which would otherwise have been unlawful. For example, in a patent license, the patentee (licensor) authorizes the licensee to exercise defined rights over the patent. The effect is to give to the licensee a right to do what he/she would otherwise be prohibited from doing, i.e., a license makes lawful what otherwise would be unlawful.[ 12 ]

The licensor may also license ‘know-how’ pertaining to the execution of the licensed patent right such as information, process, or device occurring or utilized in a business activity can also be included along with the patent right in a license agreement. Some examples of know-how are:

• (i) technical information such as formulae, techniques, and operating procedures and
• (ii) commercial information such as customer lists and sales data, marketing, professional and management procedures.

Indeed, any technical, trade, commercial, or other information, may be capable of being the subject of protection.[ 13 ]

Benefits to the licensor:

• (i) Opens new markets
• (ii) Creates new areas for revenue generation
• (iii) Helps overcome the challenge of establishing the technology in different markets especially in foreign countries – lower costs and risk and savings on distribution and marketing expenses

Benefits to the licensee are:

• (i) Savings on R&D and elimination of risks associated with R&D
• (ii) Quick exploitation of market requirements before the market interest wanes
• (iii) Ensures that products are the latest

The Role of Patent Cooperation Treaty

The patent cooperation treaty (PCT) is a multilateral treaty entered into force in 1978. Through PCT, an inventor of a member country contracting state of PCT can simultaneously obtain priority for his/her invention in all or any of the member countries, without having to file a separate application in the countries of interest, by designating them in the PCT application. All activities related to PCT are coordinated by the world intellectual property organization (WIPO) situated in Geneva.[ 14 ]

In order to protect invention in other countries, it is required to file an independent patent application in each country of interest; in some cases, within a stipulated time to obtain priority in these countries. This would entail a large investment, within a short time, to meet costs towards filing fees, translation, attorney charges, etc. In addition, it is assumed that due to the short time available for making the decision on whether to file a patent application in a country or not, may not be well founded.[ 15 ]

Inventors of contracting states of PCT on the other hand can simultaneously obtain priority for their inventions without having to file separate application in the countries of interest; thus, saving the initial investments towards filing fees, translation, etc. In addition, the system provides much longer time for filing patent application in the member countries.[ 15 , 16 ]

The time available under Paris convention for securing priority in other countries is 12 months from the date of initial filing. Under the PCT, the time available could be as much as minimum 20 and maximum 31 months. Further, an inventor is also benefited by the search report prepared under the PCT system to be sure that the claimed invention is novel. The inventor could also opt for preliminary examination before filing in other countries to be doubly sure about the patentability of the invention.[ 16 ]

Management of Intellectual Property in Pharmaceutical Industries

More than any other technological area, drugs and pharmaceuticals match the description of globalization and need to have a strong IP system most closely. Knowing that the cost of introducing a new drug into the market may cost a company anywhere between $ 300 million to $1000 million along with all the associated risks at the developmental stage, no company will like to risk its IP becoming a public property without adequate returns. Creating, obtaining, protecting, and managing IP must become a corporate activity in the same manner as the raising of resources and funds. The knowledge revolution, which we are sure to witness, will demand a special pedestal for IP and treatment in the overall decision-making process.[ 17 ]

Competition in the global pharmaceutical industry is driven by scientific knowledge rather than manufacturing know-how and a company's success will be largely dependent on its R&D efforts. Therefore, investments in R&D in the drug industry are very high as a percentage of total sales; reports suggest that it could be as much as 15% of the sale. One of the key issues in this industry is the management of innovative risks while one strives to gain a competitive advantage over rival organizations. There is high cost attached to the risk of failure in pharmaceutical R&D with the development of potential medicines that are unable to meet the stringent safety standards, being terminated, sometimes after many years of investment. For those medicines that do clear development hurdles, it takes about 8-10 years from the date when the compound was first synthesized. As product patents emerge as the main tools for protecting IP, the drug companies will have to shift their focus of R&D from development of new processes for producing known drugs towards development of a new drug molecule and new chemical entity (NCE). During the 1980s, after a period of successfully treating many diseases of short-term duration, the R&D focus shifted to long duration (chronic) diseases. While looking for the global market, one has to ensure that requirements different regulatory authorities must be satisfied.[ 18 ]

It is understood that the documents to be submitted to regulatory authorities have almost tripled in the last ten years. In addition, regulatory authorities now take much longer to approve a new drug. Consequently, the period of patent protection is reduced, resulting in the need of putting in extra efforts to earn enough profits. The situation may be more severe in the case of drugs developed through the biotechnology route especially those involving utilization of genes. It is likely that the industrialized world would soon start canvassing for longer protection for drugs. It is also possible that many governments would exercise more and more price control to meet public goals. This would on one hand emphasize the need for reduced cost of drug development, production, and marketing, and on the other hand, necessitate planning for lower profit margins so as to recover costs over a longer period. It is thus obvious that the drug industry has to wade through many conflicting requirements. Many different strategies have been evolved during the last 10 to 15 years for cost containment and trade advantage. Some of these are out sourcing of R&D activity, forming R&D partnerships and establishing strategic alliances.[ 19 ]

Nature of Pharmaceutical Industry

The race to unlock the secrets of human genome has produced an explosion of scientific knowledge and spurred the development of new technologies that are altering the economics of drug development. Biopharmaceuticals are likely to enjoy a special place and the ultimate goal will be to have personalized medicines, as everyone will have their own genome mapped and stored in a chip. Doctors will look at the information in the chip(s) and prescribe accordingly. The important IP issue associated would be the protection of such databases of personal information. Biotechnologically developed drugs will find more and more entry into the market. The protection procedure for such drug will be a little different from those conventional drugs, which are not biotechnologically developed. Microbial strains used for developing a drug or vaccine needs to be specified in the patent document. If the strain is already known and reported in the literature usually consulted by scientists, then the situation is simple. However, many new strains are discovered and developed continuously and these are deposited with International depository authorities under the Budapest Treaty. While doing a novelty search, the databases of these depositories should also be consulted. Companies do not usually go for publishing their work, but it is good to make it a practice not to disclose the invention through publications or seminars until a patent application has been filed.[ 20 ]

While dealing with microbiological inventions, it is essential to deposit the strain in one of the recognized depositories who would give a registration number to the strain which should be quoted in the patent specification. This obviates the need of describing a life form on paper. Depositing a strain also costs money, but this is not much if one is not dealing with, for example cell lines. Further, for inventions involving genes, gene expression, DNA, and RNA, the sequences also have to be described in the patent specification as has been seen in the past. The alliances could be for many different objectives such as for sharing R&D expertise and facilities, utilizing marketing networks and sharing production facilities. While entering into an R&D alliance, it is always advisable to enter into a formal agreement covering issues like ownership of IP in different countries, sharing of costs of obtaining and maintaining IP and revenue accruing from it, methods of keeping trade secrets, accounting for IP of each company before the alliance and IP created during the project but not addressed in the plan, dispute settlements. It must be remembered that an alliance would be favorable if the IP portfolio is stronger than that of concerned partner. There could be many other elements of this agreement. Many drug companies will soon use the services of academic institutions, private R&D agencies, R&D institutions under government in India and abroad by way of contract research. All the above aspects mentioned above will be useful. Special attention will have to be paid towards maintaining confidentiality of research.[ 1 – 18 ]

The current state of the pharmaceutical industry indicates that IPR are being unjustifiably strengthened and abused at the expense of competition and consumer welfare. The lack of risk and innovation on the part of the drug industry underscores the inequity that is occurring at the expense of public good. It is an unfairness that cannot be cured by legislative reform alone. While congressional efforts to close loopholes in current statutes, along with new legislation to curtail additionally unfavorable business practices of the pharmaceutical industry, may provide some mitigation, antitrust law must appropriately step in.[ 21 ] While antitrust laws have appropriately scrutinized certain business practices employed by the pharmaceutical industry, such as mergers and acquisitions and agreements not to compete, there are several other practices that need to be addressed. The grant of patents on minor elements of an old drug, reformulations of old drugs to secure new patents, and the use of advertising and brand name development to increase the barriers for generic market entrants are all areas in which antitrust law can help stabilize the balance between rewarding innovation and preserving competition.[ 20 ]

Traditional medicine dealing with natural botanical products is an important part of human health care in many developing countries and also in developed countries, increasing their commercial value. The world market for such medicines has reached US $ 60 billion, with annual growth rates of between 5% and 15%. Although purely traditional knowledge based medicines do not qualify for patent, people often claim so. Researchers or companies may also claim IPR over biological resources and/or traditional knowledge, after slightly modifying them. The fast growth of patent applications related to herbal medicine shows this trend clearly. The patent applications in the field of natural products, traditional herbal medicine and herbal medicinal products are dealt with own IPR policies of each country as food, pharmaceutical and cosmetics purview, whichever appropriate. Medicinal plants and related plant products are important targets of patent claims since they have become of great interest to the global organized herbal drug and cosmetic industries.[ 22 ]

Some Special Aspects of Drug Patent Specification

Writing patent specification is a highly professional skill, which is acquired over a period of time and needs a good combination of scientific, technological, and legal knowledge. Claims in any patent specification constitute the soul of the patent over which legal proprietary is sought. Discovery of a new property in a known material is not patentable. If one can put the property to a practical use one has made an invention which may be patentable. A discovery that a known substance is able to withstand mechanical shock would not be patentable but a railway sleeper made from the material could well be patented. A substance may not be new but has been found to have a new property. It may be possible to patent it in combination with some other known substances if in combination they exhibit some new result. The reason is that no one has earlier used that combination for producing an insecticide or fertilizer or drug. It is quite possible that an inventor has created a new molecule but its precise structure is not known. In such a case, description of the substance along with its properties and the method of producing the same will play an important role.[ 23 ]

Combination of known substances into useful products may be a subject matter of a patent if the substances have some working relationship when combined together. In this case, no chemical reaction takes place. It confers only a limited protection. Any use by others of individual parts of the combination is beyond the scope of the patent. For example, a patent on aqua regia will not prohibit any one from mixing the two acids in different proportions and obtaining new patents. Methods of treatment for humans and animals are not patentable in most of the countries (one exception is USA) as they are not considered capable of industrial application. In case of new pharmaceutical use of a known substance, one should be careful in writing claims as the claim should not give an impression of a method of treatment. Most of the applications relate to drugs and pharmaceuticals including herbal drugs. A limited number of applications relate to engineering, electronics, and chemicals. About 62% of the applications are related to drugs and pharmaceuticals.[ 1 – 24 ]

CONCLUSIONS

It is obvious that management of IP and IPR is a multidimensional task and calls for many different actions and strategies which need to be aligned with national laws and international treaties and practices. It is no longer driven purely by a national perspective. IP and its associated rights are seriously influenced by the market needs, market response, cost involved in translating IP into commercial venture and so on. In other words, trade and commerce considerations are important in the management of IPR. Different forms of IPR demand different treatment, handling, planning, and strategies and engagement of persons with different domain knowledge such as science, engineering, medicines, law, finance, marketing, and economics. Each industry should evolve its own IP policies, management style, strategies, etc. depending on its area of specialty. Pharmaceutical industry currently has an evolving IP strategy. Since there exists the increased possibility that some IPR are invalid, antitrust law, therefore, needs to step in to ensure that invalid rights are not being unlawfully asserted to establish and maintain illegitimate, albeit limited, monopolies within the pharmaceutical industry. Still many things remain to be resolved in this context.

Source of Support: Nil

Conflict of Interest: Nil.

• [email protected]

INDIAN JOURNAL OF CORPORATE LAW AND POLICY

• Individual Publications
• Blog Submission Guidelines

Student, Institute of Company Secretaries of India

View all posts

Artificial Intelligence (AI) and Intellectual Property Rights (IPR)- Legal status and the future

• February 15, 2022

The world is constantly evolving, and we have witnessed some major technical revolutions and some great technological ideas in the age of pandemics. Assessment of the efficacy of technical innovations from a legal perspective is hence, essential. This piece will talk about the possible metamorphosis of artificial intelligence into the world of Intellectual Property. Since Artificial Intelligence (AI) is evolving at such an instantaneous speed, this entreats in-depth scrutiny of the existing IP laws.

As of now, there are no provisions associated with Artificial Intelligence (AI) in the course of Intellectual Property. However, there lies a substantial amount of involvement and correlation between the two that we will discuss later in this article. Presently, Artificial Intelligence (AI) is going through extensive advancement in all parts of the world and has been coinciding with the ambit of Intellectual Property Rights (IPR). Henceforth, there is a dire need of acknowledging it to the system and amending some of the existing norms. 

Introduction

The history of the word ‘Artificial Intelligence’ can be traced back to the year 1956 when it was first coined at Dartmouth college. [1]

Artificial Intelligence (AI) is essentially the intelligence produced by ‘machines’. For a clear and more sophisticated understanding, we can say that any intelligence that is opposed to the natural flow of intelligence demonstrated by ‘animals and living beings’ falls under the connotation of Artificial Intelligence.

The Intellectual Property (IP) is affiliated to any authentic invention of human intelligence such as artistic, literary, technical, or scientific creation. Primarily, these are the intangible properties that existed in the creator’s thought (idea) which is later converted to tangible (existing in reality) property. On the other hand, Intellectual Property Rights (IPR) refer to the legal rights granted to the creator or founder to safeguard his invention or design for a certain period. IPR in India is rather new and still in its budding stage.

Today there are multiple examples of how human inventions and robotics evolved out of human intellect are continuously working towards creating new things and evolving new ideas out of their algorithm that is valuable for day-to-day life and these things are constantly making our life more comfortable. Discussions on discerning the precise legal entity concerning the ownership of rights for such innovations have reached an impasse.

Intellectual Property (IP) and Artificial Intelligence (AI)

The Intellectual Property Rights (IPR) is a vital tool to safeguard and incentivize the innovations of the human intellect. Artificial intelligence has turned out to be a relatively new area of debate concerning laws such as; copyright and patent laws. Differentiating between actual human consciousness and artificial consciousness is often the central area of discussions involving Intellectual Property Rights (IPR) and Artificial Intelligence (AI). One of the prominent predicaments resides in determining the liability in cases of failures of such innovations. The WIPO (World Intellectual Property Organization) is constantly involved in discourses and actively trying to find solutions to set aside such problems.

The existing Intellectual Property (IP) laws are not competent to address issues regarding the identification of inventors and other violations when Artificial Intelligence (AI) is involved with creation. There are numerous challenges before the policymakers, and this has been a topic of continuous debate in the domain of lawmakers and experts.

World: The Contemporary Legal Scenario

Presently, there is no specific law governing the role of self-involvement of Artificial Intelligence (AI) in innovations. However, there have been some legal developments regarding the topic over time. The United States consider human as copyright holders. The situation in the United States is also a challenging one. Fairly recently, The United States Patent and Trademark Office (USPTO) declined a petition involving the Artificial Intelligence (AI) systems and inventors. [2]

DABUS (an AI system that stands for “device for the autonomous bootstrapping of unified sentience”) created by Stephen Thaler has a long history in different jurisdictions. [3] Some of them are ongoing as well. The European Union maintains a similar stance as the United States. Many Patent courts have repeatedly declined to give the inventor status to Artificial Intelligence (AI) systems. But recently, South Africa has become the first country to grant patent status to DABUS. [4] Although, there are still some fusses regarding the decision around the world of experts. However, this shows the potential of AI systems and their integration with Intellectual Property (IP). Even the Australian courts have recently found out that Artificial Intelligence (AI) is capable of being an ‘inventor’. [5]

Other than these countries, we have seen Japan being highly involved with the workings of the Artificial Intelligence (AI) systems and their possible future as evident from the ‘AI Strategy 2019 AI for Everyone- People, Industries, Regions and Governments (2019)’.

Several countries and patent offices are taking cognizance of the development of Artificial Intelligence (AI) systems, however, the majority of them have not been able to turn things around.

India – Ownership and Artificial Intelligence (AI)

India is one of the major countries when we talk about technological advancement. India has an enormous population. And with such a tremendous population, there is immense commercial scope for the advent of tech companies in the country. Moreover, India is still in its developing phase. There is an adequate establishment of copyright and patent laws in the legal framework of the country. But, like many other countries, India also lacks a provision for the regulation of Artificial Intelligence (AI) with the Intellectual Property Rights (IPR).

The concept of giving an inventor status to machines is still questionable and unfamiliar in the country. This comes from the implied and direct assumptions from the Copyright Act and the Patents Act of 1970, respectively.

Some of the provisions in the existing laws restrict the expansion of the idea of a creator. Thus, only humans can obtain protection under the existing laws in India.

There are a lot of challenges wherever Intellectual Property (IP) crosses the path of Artificial Intelligence (AI) established innovations regarding disclosure, copyright laws, definitions of inventor and owner, and violations. The current model in the majority of the world is not equipped well enough to answer such questions. Artificial Intelligence (AI) is going through monumental growth, and with an increase in complexity of the systems, the existing laws are unable to cope-up with the rapid pace of technological developments. However, the increasing cognition and new disclosures have reinvigorated this challenging process. For the time being, the conventional laws need to acknowledge the technology as soon as possible because the world will keep on developing and becoming more and more complex. There are organizations like WIPO (World Intellectual Property Organization) that are holding conversations to facilitate the Intellectual Property (IP) laws with the complex nature of artificial intelligence (AI). There is a noticeable demand for the formulation of Intellectual Property (IP) laws that can protect the products of artificial intelligence (AI) and machine innovation so technology can move forward.

[1] Rockwell Anyoha, The History of Artificial Intelligence, SCIENCE IN THE NEWS, HARVARD UNIVERSITY (Aug. 28, 2017), https://sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence/

[2] Tyler Sonnemaker, No, an artificial intelligence can’t legally invent something – only ‘natural persons’ can, says US patent office, BUSINESS INSIDER (Apr 30. 2020, 1:11 AM), https://www.businessinsider.com/artificial-inteligence-cant-legally-named-inventor-us-patent-office-ruling-2020-4?r=US&IR=T

[3] Miguel Bibe, DABUS: the ‘natural person’ problem, INVENTA (Sep. 27, 2021), https://inventa.com/en/news/article/681/dabus-the-natural-person-problem

[4] Meshandren Naidoo, In a world first. South Africa grants a patent to an artificial intelligence system, QUARTZ MEMBERSHIP (Aug. 9, 2021), https://qz.com/africa/2044477/south-africa-grants-patent-to-an-ai-system-known-as-dabus/

[5] Rebecca Currey & Jane Owen, In the Courts: Australian Court finds AI systems can be “inventors”, WIPO INTERNATIONAL (Sept 2021), https://www.wipo.int/wipo_magazine/en/2021/03/article_0006.html

More to explorer

INTELLECTUAL PROPERTY RIGHTS & THEIR SIGNIFICANCE IN CONTEXT WITH ARTIFICIAL INTELLIGENCE

ABSTRACT Living in a connected world, working in a digital, mobile manner using cloud computing services, for instance, is still very new

CROSS BORDER INSOLVENCY- STATE BANK OF INDIA v. JET AIRWAYS (INDIA) LTD

Abstract Globalisation has established a high degree of interdependence between nations and their respective companies. Cross-border insolvency is a matter of immense

Perpetual Succession — A Detailed Study

Abstract This paper is a  study of one of the main topics related to Company Law Perpetual Succession. In this article one

Indian Journal of Corporate Law and Policy

Indian Journal of Corporate Law and Policy is a double-blind peer-reviewed, Triannual Journal.

 It is published by the Society for Progress in Research, Education and Development in Law (SPRED LAW).

Copyright Form

Related links.

• Mail : [email protected]

©2020. IJCLP. All Rights Reserved.

• Published on May 31, 2017
• In Endless Origins

10 most impressive Research Papers around Artificial Intelligence

• By Amit Paul Chowdhury

Artificial Intelligence research advances are transforming technology as we know it. The AI research community is solving some of the most technology problems related to software and hardware infrastructure, theory and algorithms. Interestingly, the field of AI AI research has drawn acolytes from the non-tech field as well. Case in point — prolific Hollywood actor Kristen Stewart’s highly publicized paper on Artificial Intelligence, originally published at Cornell University library’s open access site .  Stewart co-authored the paper , titled “ Bringing Impressionism to Life with Neural Style Transfer in Come Swim ” with the American poet and literary critic David Shapiro and Adobe Research Engineer Bhautik Joshi .

Essentially, the AI-based paper talks about the style transfer techniques used in her short film Come Swim . However, Stewart’s detractors dismissed it as another “high-level case study.”

Meanwhile, the community is awash with ground-breaking research papers around AI.   Analytics India Magazine lists down the most cited scientific papers around AI, machine intelligence, and computer vision , that will give a perspective on the technology and its applications.

Most of these papers have been chosen on the basis of citation value for each. Some of these papers take into account a Highly Influential Citation count (HIC) and Citation Velocity (CV). Citation Velocity is the weighted average number of citations per year over the last 3 years.

A Computational Approach to Edge Detection : Originally published in 1986 and authored by John Canny this paper, on the computational approach to edge detection, has approximately 9724 citations . The success of this approach is defined by a comprehensive set of goals for the computation of edge points. These goals must be precise enough to delimit the desired behavior of the detector while making minimal assumptions about the form of the solution.

Besides, the paper also presents a general method, called feature synthesis, for the fine-to-coarse integration of information from operators at different scales. This helps in establishing the fact that edge detector performance improves considerably as the operator point spread function is extended along the edge.

A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence : This research paper was co-written by John McCarthy, Marvin L. Minsky, Nathaniel Rochester, Claude E. Shannon, and published in the year 1955. This summer research proposal defined the field, and has another first to its name — it is the first paper to use the term Artificial Intelligence. The proposal invited researchers to the Dartmouth conference , which is widely considered the “birth of AI”.

A Threshold Selection Method from Gray-Level Histograms : The paper was authored by Nobuyuki Otsu and published in 1979 . It has received 7849 paper citations so far. Through this paper, Otsu discusses a nonparametric and unsupervised method of automatic threshold selection for picture segmentation.

The paper delves into how an optimal threshold is selected by the discriminant criterion to maximize the separability of the resultant classes in gray levels. The procedure utilizes only the zeroth- and first-order cumulative moments of the gray-level histogram. The method can be easily applied across multi threshold problems. The paper validates the method by presenting several experimental results.

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift : This 2015 article was co-written by Sergey Ioffe and Christian Szegedy . The paper received 946 citations and reflects on a HIC score of 56.

The paper talks about how training deep neural networks is complicated by the fact that the distribution of each layer’s inputs changes during training. This is a result of change in parameters of the previous layers. The phenomenon is termed as internal covariate shift. This issue is addressed by normalizing layer inputs.

Batch normalization achieves the same accuracy with 14 times fewer training steps when applied to a state-of-the-art image classification model. In other words, Batch Normalization beats the original model by a significant margin.

Deep Residual Learning for Image Recognition : The 2016 paper was co-authored by Kaiming He, Xiangyu Zhang, and Shaoqing Ren. The paper has been cited 1436 times, reflecting on a HIC value of 137 and a CV of 582 . The authors have delved into residual learning framework to ease the training of deep neural networks that are substantially deeper than those used previously.

Besides, the research paper explicitly reformulates the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. The research also delves into how comprehensive empirical evidence show that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth.

Distinctive Image Features from Scale-Invariant Keypoints : This article was authored by David G. Lowe in 2004 . The paper received 21528 citations  and explores the method for extracting distinctive invariant features from images. These can be utilized to perform reliable matching between different views of an object or scene. The features are invariant to image scale and rotation, and are shown to provide robust matching across a substantial range of affine distortion, change in 3D viewpoint, addition of noise, and change in illumination.

The paper additionally delves into an approach which leverages these features for image recognition. This approach can help identify objects among clutter and occlusion while achieving near real-time performance.

Dropout: a simple way to prevent neural networks from overfitting : The 2014 paper was co-authored by Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov . The paper has been cited around 2084 times , with a HIC and CV value of 142 and 536 respectively . Deep neural nets with a large number of parameters are very powerful machine learning systems. However, overfitting is a serious problem in such networks.

The central premise of the paper is to drop units (along with their connections) from the neural network during training, thus preventing units from co-adapting too much. This helps in significantly reducing overfitting, while furnishing major improvements over other regularization methods.

Induction of decision trees : Authored by J. R. Quinlan , this scientific paper was originally published in 1986 and summarizes an approach to synthesizing decision trees that has been used in a variety of systems. The paper specifically describes one such system, ID3, in detail. Additionally, the paper discusses a reported shortcoming of the basic algorithm , besides comparing the two methods of overcoming it. To conclude the paper, the author presents illustrations of current research directions.

Large-Scale Video Classification with Convolutional Neural Networks : This 2014 paper was co-written by 6 authors, Andrej Karpathy, George Toderici, Sanketh Shetty, Thomas Leung, Rahul Sukthankar, and Li Fei-Fei. The paper has been cited over 865 times , and reflects on a HIC score of 24 , and a CV of 239 .

Convolutional Neural Networks (CNNs) are proven to stand as a powerful class of models for image recognition problems. These results encouraged the authors to provide an extensive empirical evaluation of CNNs on large-scale video classification. This was accomplished using a new dataset of 1 million YouTube videos belonging to 487 classes.

Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference : The paper was published in 1988 . Judea Pearl is the author to this article. The paper presents a complete and accessible account of the theoretical foundations and computational methods that underlie plausible reasoning under uncertainty.

Pearl furnishes a provides a coherent explication of probability as a language for reasoning with partial belief and offers a unifying perspective on other AI approaches to uncertainty, such as the Dempster-Shafer formalism , truth maintenance systems, and nonmonotonic logic.

Download our Mobile App

When Sam Met Ollie 

Russian Tech Goliath Yandex Open Sources DataLens

AI Trust Grows When Hidden

Why Google will Not Rank Websites Just Based on SEOs 

Why Microsoft is Using Nuclear to Fuel Its Data Centers? 

MIT’s New AI-Powered Co-Pilot Will Redefine Aviation Safety

Our upcoming events, 3 ways to join our community, telegram group.

Discover special offers, top stories, upcoming events, and more.

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Subscribe to our Daily newsletter

Get our daily awesome stories & videos in your inbox, most popular.

KPI Partners: Accelerating Business Growth with Modernization and Monetization of Data Assets

KPI Partners, a leader in digital transformation, is dedicated to accelerate business profitability by providing customized data analytics and digital transformation services and solutions to address the unique needs of each client. 

How Generative AI is Revolutionising Data Science Tools

Einblick Prompt enables users to create complete data workflows using natural language, accelerating various stages of data science and analytics. Einblick has effectively combined the capabilities of a Jupyter notebook with the user-friendliness of ChatGPT.

Google is Officially Killing the Internet with AI

While Google has long preached ‘helpful content written by people, for people, in search results’, its recent actions suggest otherwise

Why Neeva Pivoted to Enterprise AI with Snowflake 

A part of the founder’s job is to see beyond the corner and get a feel of what is coming

Larry Lulls Everyone into Generative AI

“Does that mean we are going to have massive layoffs at Oracle? No, we are too ambitious for that,” said Larry Ellison

[Report] LLM Economics – A Guide to Generative AI Implementation Cost

This will serve as a vital tool helping stakeholders to assess the potential costs and benefits associated with different implementation strategies, whether it be through API or open-source pathways.

OpenAI Banks on Red Team to Win the Moral Battle

OpenAI’s latest announcement on inviting people for forming a Red Teaming Network may not be a completely new initiative. Then, why now?

The Proportional Rise of Inference and CPUs

As inference costs scale with the number of users, the impending shift in AI landscape has forced NVIDIA, AMD, Intel to pay heed to it

Intel Goes All in on AI

Pat Gelsinger said, there are three types of chip manufacturers, “you’re big, you’re niche or you’re dead”

OpenAI Wins Again 

Puts Gemini and Midjourney in a tough spot

Our mission is to bring about better-informed and more conscious decisions about technology through authoritative, influential, and trustworthy journalism.

Shape the future of tech.

© Analytics India Magazine Pvt Ltd & AIM Media House LLC 2023

• Terms of use
• Privacy Policy

• Open Access
• Published: 22 September 2023

Revolutionizing healthcare: the role of artificial intelligence in clinical practice

• Shuroug A. Alowais 1 , 2 , 3 ,
• Sahar S. Alghamdi 2 , 3 , 4 ,
• Nada Alsuhebany 1 , 2 , 3 ,
• Tariq Alqahtani 2 , 3 , 4 ,
• Abdulrahman I. Alshaya 1 , 2 , 3 ,
• Sumaya N. Almohareb 1 , 2 , 3 ,
• Atheer Aldairem 1 , 2 , 3 ,
• Mohammed Alrashed 1 , 2 , 3 ,
• Khalid Bin Saleh 1 , 2 , 3 ,
• Hisham A. Badreldin 1 , 2 , 3 ,
• Majed S. Al Yami 1 , 2 , 3 ,
• Shmeylan Al Harbi 1 , 2 , 3 &
• Abdulkareem M. Albekairy 1 , 2 , 3  

BMC Medical Education volume  23 , Article number:  689 ( 2023 ) Cite this article

848 Accesses

19 Altmetric

Metrics details

Introduction

Healthcare systems are complex and challenging for all stakeholders, but artificial intelligence (AI) has transformed various fields, including healthcare, with the potential to improve patient care and quality of life. Rapid AI advancements can revolutionize healthcare by integrating it into clinical practice. Reporting AI’s role in clinical practice is crucial for successful implementation by equipping healthcare providers with essential knowledge and tools.

Research Significance

This review article provides a comprehensive and up-to-date overview of the current state of AI in clinical practice, including its potential applications in disease diagnosis, treatment recommendations, and patient engagement. It also discusses the associated challenges, covering ethical and legal considerations and the need for human expertise. By doing so, it enhances understanding of AI’s significance in healthcare and supports healthcare organizations in effectively adopting AI technologies.

Materials and Methods

The current investigation analyzed the use of AI in the healthcare system with a comprehensive review of relevant indexed literature, such as PubMed/Medline, Scopus, and EMBASE, with no time constraints but limited to articles published in English. The focused question explores the impact of applying AI in healthcare settings and the potential outcomes of this application.

Integrating AI into healthcare holds excellent potential for improving disease diagnosis, treatment selection, and clinical laboratory testing. AI tools can leverage large datasets and identify patterns to surpass human performance in several healthcare aspects. AI offers increased accuracy, reduced costs, and time savings while minimizing human errors. It can revolutionize personalized medicine, optimize medication dosages, enhance population health management, establish guidelines, provide virtual health assistants, support mental health care, improve patient education, and influence patient-physician trust.

AI can be used to diagnose diseases, develop personalized treatment plans, and assist clinicians with decision-making. Rather than simply automating tasks, AI is about developing technologies that can enhance patient care across healthcare settings. However, challenges related to data privacy, bias, and the need for human expertise must be addressed for the responsible and effective implementation of AI in healthcare.

Peer Review reports

Artificial Intelligence (AI) is a rapidly evolving field of computer science that aims to create machines that can perform tasks that typically require human intelligence. AI includes various techniques such as machine learning (ML), deep learning (DL), and natural language processing (NLP). Large Language Models (LLMs) are a type of AI algorithm that uses deep learning techniques and massively large data sets to understand, summarize, generate, and predict new text-based content [ 1 , 2 , 3 ]. LLMs have been architected to generate text-based content and possess broad applicability for various NLP tasks, including text generation, translation, content summary, rewriting, classification, categorization, and sentiment analysis. NLP is a subfield of AI that focuses on the interaction between computers and humans through natural language, including understanding, interpreting, and generating human language. NLP involves various techniques such as text mining, sentiment analysis, speech recognition, and machine translation. Over the years, AI has undergone significant transformations, from the early days of rule-based systems to the current era of ML and deep learning algorithms [ 1 , 2 , 3 ].

AI has evolved since the first AI program was developed in 1951 by Christopher Strachey. At that time, AI was in its infancy and was primarily an academic research topic. In 1956, John McCarthy organized the Dartmouth Conference, where he coined the term “Artificial Intelligence.“ This event marked the beginning of the modern AI era. In the 1960 and 1970 s, AI research focused on rule-based and expert systems. However, this approach was limited by the need for more computing power and data [ 4 ].

In the 1980 and 1990 s, AI research shifted to ML and neural networks, which allowed machines to learn from data and improve their performance over time. This period saw the development of systems such as IBM’s Deep Blue, which defeated world chess champion Garry Kasparov in 1997. In the 2000s, AI research continued to evolve, focusing on NLP and computer vision, which led to the development of virtual assistants, such as Apple’s Siri and Amazon’s Alexa, which could understand natural language and respond to user requests (Fig.  1 ) [ 3 , 4 ].

Tracing the Evolution of AI with a Better Understanding of the Relationship Between AI, ML, DL, and NLP

Today, AI is transforming healthcare, finance, and transportation, among other fields, and its impact is only set to grow. In academia, AI has been used to develop intelligent tutoring systems, which are computer programs that can adapt to the needs of individual students. These systems have improved student learning outcomes in various subjects, including math and science. In research, AI has been used to analyze large datasets and identify patterns that would be difficult for humans to detect; this has led to breakthroughs in fields such as genomics and drug discovery. AI has been used in healthcare settings to develop diagnostic tools and personalized treatment plans. As AI continues to evolve, it is crucial to ensure that it is developed responsibly and for the benefit of all [ 5 , 6 , 7 , 8 ].

The rapid progression of AI technology presents an opportunity for its application in clinical practice, potentially revolutionizing healthcare services. It is imperative to document and disseminate information regarding AI’s role in clinical practice, to equip healthcare providers with the knowledge and tools necessary for effective implementation in patient care. This review article aims to explore the current state of AI in healthcare, its potential benefits, limitations, and challenges, and to provide insights into its future development. By doing so, this review aims to contribute to a better understanding of AI’s role in healthcare and facilitate its integration into clinical practice.

Materials and methods

Search strategy and inclusion.

Indexed databases, including PubMed/Medline (National Library of Medicine), Scopus, and EMBASE, were independently searched with notime restrictions, but the searches were limited to the English language.

Databases search protocol and keywords

In the review article, the authors extensively examined the use of AI in healthcare settings. The authors analyzed various combinations of keywords such as NLP in healthcare, ML in healthcare, DL in healthcare, LLM in healthcare, AI in personalized medicine, AI in patient monitoring, AI ethics in healthcare, predictive analytics in healthcare, AI in medical diagnosis, and AI applications in healthcare. By imposing language restrictions, the authors ensured a comprehensive analysis of the topic.

Data extraction

Publications were screened through a meticulous review of titles and abstracts. Only those that met the specific criteria were included. Any disagreements or concerns about the literature or methodology were discussed in detail among the authors.

AI assistance in diagnostics

Diagnosis accuracy.

With all the advances in medicine, effective disease diagnosis is still considered a challenge on a global scale. The development of early diagnostic tools is an ongoing challenge due to the complexity of the various disease mechanisms and the underlying symptoms. AI can revolutionize different aspects of health care, including diagnosis. ML is an area of AI that uses data as an input resource in which the accuracy is highly dependent on the quantity as well as the quality of the input data that can combat some of the challenges and complexity of diagnosis [ 9 ]. ML, in short, can assist in decision-making, manage workflow, and automate tasks in a timely and cost-effective manner. Also, deep learning added layers utilizing Convolutional Neural Networks (CNN) and data mining techniques that help identify data patterns. These are highly applicable in identifying key disease detection patterns among big datasets. These tools are highly applicable in healthcare systems for diagnosing, predicting, or classifying diseases [ 10 ].

AI is still in its early stages of being fully utilized for medical diagnosis. However, more data are emerging for the application of AI in diagnosing different diseases, such as cancer. A study was published in the UK where authors input a large dataset of mammograms into an AI system for breast cancer diagnosis. This study showed that utilizing an AI system to interpret mammograms had an absolute reduction in false positives and false negatives by 5.7% and 9.4%, respectively [ 11 ]. Another study was conducted in South Korea, where authors compared AI diagnoses of breast cancer versus radiologists. The AI-utilized diagnosis was more sensitive to diagnose breast cancer with mass compared to radiologists, 90% vs. 78%, respectively. Also, AI was better at detecting early breast cancer (91%) than radiologists 74% [ 12 ].

Furthermore, a study utilized deep learning to detect skin cancer which showed that an AI using CNN accurately diagnosed melanoma cases compared to dermatologists and recommended treatment options [ 13 , 14 ]. Researchers utilized AI technology in many other disease states, such as detecting diabetic retinopathy [ 15 ] and EKG abnormality and predicting risk factors for cardiovascular diseases [ 16 , 17 ]. Furthermore, deep learning algorithms are used to detect pneumonia from chest radiography with sensitivity and specificity of 96% and 64% compared to radiologists 50% and 73%, respectively [ 18 ]. Also, a study was done on a dataset of 625 cases to diagnose acute appendicitis early to predict the need for appendix surgery using various ML techniques; the results showed that the random forest algorithm achieved the highest performance, accurately predicting appendicitis in 83.75% of cases, with a precision of 84.11%, sensitivity of 81.08%, and specificity of 81.01%. The improved method aids healthcare specialists in making informed decisions for appendicitis diagnoses and treatment. Furthermore, the authors suggest that similar techniques can be utilized to analyze images of patients with appendicitis or even to detect infections such as COVID-19 using blood specimens or images [ 19 ].

AI tools can improve accuracy, reduce costs, and save time compared to traditional diagnostic methods. Additionally, AI can reduce the risk of human errors and provide more accurate results in less time. In the future, AI technology could be used to support medical decisions by providing clinicians with real-time assistance and insights. Researchers continue exploring ways to use AI in medical diagnosis and treatment, such as analyzing medical images, X-rays, CT scans, and MRIs. By leveraging ML techniques, AI can also help identify abnormalities, detect fractures, tumors, or other conditions, and provide quantitative measurements for faster and more accurate medical diagnosis.

Clinical laboratory testing provides critical information for diagnosing, treating, and monitoring diseases. It is an essential part of modern healthcare which continuously incorporates new technology to support clinical decision-making and patient safety [ 20 ]. AI has the potential to transform clinical laboratory testing by improving the accuracy, speed, and efficiency of laboratory processes. The role of AI in clinical microbiology is currently progressing and expanding. Several ML systems were developed to detect, identify, and quantify microorganisms, diagnose and classify diseases, and predict clinical outcomes. These ML systems used data from various sources to build the AI diagnosis such as genomic data of microorganisms, gene sequencing, metagenomic sequencing results of the original specimen, and microscopic imaging [ 21 ]. Moreover, gram stain classification to gram positives/negatives and cocci/rods is another essential application of using deep convolutional neural networks that reveal high sensitivity and specificity [ 22 ]. A published systematic review showed that numerous MLs were evaluated for microorganism identification and antibiotic susceptibility testing; however, several limitations are associated with the current models that must be addressed before incorporating them into clinical practice [ 23 ]. For malaria, Taesik et al. found that using ML algorithms combined with digital in-line holographic microscopy (DIHM) effectively detected malaria-infected red blood cells without staining. This AI technology is rapid, sensitive, and cost-effective in diagnosing malaria [ 24 ].

The projected benefits of using AI in clinical laboratories include but are not limited to, increased efficacy and precision. Automated techniques in blood cultures, susceptibility testing, and molecular platforms have become standard in numerous laboratories globally, contributing significantly to laboratory efficiency [ 21 , 25 ]. Automation and AI have substantially improved laboratory efficiency in areas like blood cultures, susceptibility testing, and molecular platforms. This allows for a result within the first 24 to 48 h, facilitating the selection of suitable antibiotic treatment for patients with positive blood cultures [ 21 , 26 ]. Consequently, incorporating AI in clinical microbiology laboratories can assist in choosing appropriate antibiotic treatment regimens, a critical factor in achieving high cure rates for various infectious diseases [ 21 , 26 ].

ML research in medicine has rapidly expanded, which could greatly help the healthcare providers in the emergency department (ED) as they face challenging difficulties from the rising burden of diseases, greater demand for time and health services, higher societal expectations, and increasing health expenditures [ 27 ]. Emergency department providers understand that integrating AI into their work processes is necessary for solving these problems by enhancing efficiency, and accuracy, and improving patient outcomes [ 28 , 29 ]. Additionally, there may be a chance for algorithm support and automated decision-making to optimize ED flow measurements and resource allocation [ 30 ]. AI algorithms can analyze patient data to assist with triaging patients based on urgency; this helps prioritize high-risk cases, reducing waiting times and improving patient flow [ 31 ]. Introducing a reliable symptom assessment tool can rule out other causes of illness to reduce the number of unnecessary visits to the ED. A series of AI-enabled machines can directly question the patient, and a sufficient explanation is provided at the end to ensure appropriate assessment and plan.

Moreover, AI-powered decision support systems can provide real-time suggestions to healthcare providers, aiding diagnosis, and treatment decisions. Patients are evaluated in the ED with little information, and physicians frequently must weigh probabilities when risk stratifying and making decisions. Faster clinical data interpretation is crucial in ED to classify the seriousness of the situation and the need for immediate intervention. The risk of misdiagnosing patients is one of the most critical problems affecting medical practitioners and healthcare systems. Diagnostic mistakes in the healthcare sector can be expensive and fatal. A study found that diagnostic errors, particularly in patients who visit the ED, directly contribute to a greater mortality rate and a more extended hospital stay [ 32 ]. Fortunately, AI can assist in the early detection of patients with life-threatening diseases and promptly alert clinicians so the patients can receive immediate attention. Lastly, AI can help optimize health care sources in the ED by predicting patient demand, optimizing therapy selection (medication, dose, route of administration, and urgency of intervention), and suggesting emergency department length of stay. By analyzing patient-specific data, AI systems can offer insights into optimal therapy selection, improving efficiency and reducing overcrowding.

AI in genomic medicine

The fusion of AI and genotype analysis holds immense promise in the realms of disease surveillance, prediction, and personalized medicine [ 33 ]. When applied to large populations, AI can effectively monitor for emerging disease threats (such as COVID-19), while genomic data can provide valuable insights into genetic markers associated with increased susceptibility to specific diseases [ 34 ] By training ML algorithms to identify these markers in real-time data, we can facilitate the early detection of potential outbreaks. Moreover, the use of genotype data can aid in refining disease risk predictions, as ML algorithms can recognize complex patterns of genetic variations linked with disease susceptibility that might elude traditional statistical methods as summarized in Fig.  2 [ 35 , 36 ]. The prediction of phenotypes, or observable characteristics shaped by genes and environmental factors, also becomes possible with this combination.

Schematic representation of the process starting with the extraction of DNA/RNA, followed by sequencing. The subsequent genotypic alignment is performed using neural networks and deep learning. Probability calculations are achieved through applying statistical methods and M: The graph’s Y-axis denotes the probability (expressed in percentage) of a particular type of disease (hypertension, depression, breast cancer, and Alzheimer’s disease), while the X-axis signifies the count of gene mutations. Negative numbers indicate gene deletions, whereas positive values represent gene additions or nucleic acid mutations

ML algorithms make it feasible to predict a spectrum of phenotypes ranging from simple traits like eye color to more intricate ones like the response to certain medications or disease susceptibility. A specific area where AI and ML have demonstrated significant efficacy is the identification of genetic variants associated with distinctive traits or pathologies. Examining extensive genomic datasets allows these techniques to detect intricate patterns often elusive to manual analysis. For instance, a groundbreaking study employed a deep neural network to identify genetic variants associated with autism spectrum disorder (ASD), successfully predicting ASD status by relying solely on genomic data [ 37 ]. In the field of oncology, categorizing cancers into clinically relevant molecular subtypes can be accomplished using transcriptomic profiling. Such molecular classifications, first developed for breast cancer and later extended to other cancers like colorectal, ovarian, and sarcomas, hold substantial implications for diagnosis, prognosis, and treatment selection [ 38 , 39 ]. Traditional computational methods for subtyping cancers, such as support vector machines (SVMs) or k-nearest neighbors, are susceptible to errors due to batch effects and may only focus on a small set of signature genes, thus neglecting vital biological information [ 40 ].

The advent of high-throughput genomic sequencing technologies, combined with advancements in AI and ML, has laid a strong foundation for accelerating personalized medicine and drug discovery [ 41 ]. Despite being a treasure trove of valuable insights, the complex nature of extensive genomic data presents substantial obstacles to its interpretation. The field of drug discovery has dramatically benefited from the application of AI and ML. The simultaneous analysis of extensive genomic data and other clinical parameters, such as drug efficacy or adverse effects, facilitates the identification of novel therapeutic targets or the repurposing of existing drugs for new applications [ 42 , 43 , 44 , 45 , 46 ]. One of the prevalent challenges in drug development is non-clinical toxicity, which leads to a significant percentage of drug failures during clinical trials. However, the rise of computational modeling is opening up the feasibility of predicting drug toxicity, which can be instrumental in improving the drug development process [ 46 ]. This capability is particularly vital for addressing common types of drug toxicity, such as cardiotoxicity and hepatotoxicity, which often lead to post-market withdrawal of drugs.

AI assistance in treatment

Precision medicine and clinical decision support.

Personalized treatment, also known as precision medicine or personalized medicine, is an approach that tailors medical care to individual patients based on their unique characteristics, such as genetics, environment, lifestyle, and biomarkers [ 47 ]. This individualized approach aims to improve patient outcomes by providing targeted interventions that are more effective, efficient, and safe. AI has emerged as a valuable tool in advancing personalized treatment, offering the potential to analyze complex datasets, predict outcomes, and optimize treatment strategies [ 47 , 48 ]. Personalized treatment represents a pioneering field that demonstrates the potential of precision medicine on a large scale [ 49 ]. Nevertheless, the ability to provide real-time recommendations relies on the advancement of ML algorithms capable of predicting patients who may require specific medications based on genomic information. The key to tailoring medications and dosages to patients lies in the pre-emptive genotyping of patients prior to the actual need for such information [ 49 , 50 ].

The potential applications of AI in assisting clinicians with treatment decisions, particularly in predicting therapy response, have gained recognition [ 49 ]. A study conducted by Huang et al. where authors utilized patients’ gene expression data for training a support ML, successfully predicted the response to chemotherapy [ 51 ]. In this study, the authors included 175 cancer patients incorporating their gene-expression profiles to predict the patients’ responses to various standard-of-care chemotherapies. Notably, the research showed encouraging outcomes, achieving a prediction accuracy of over 80% across multiple drugs. These findings demonstrate the promising role of AI in treatment response prediction. In another study performed by Sheu et al., the authors aimed to predict the response to different classes of antidepressants using electronic health records (EHR) of 17,556 patients and AI [ 52 ]. The AI models considered features predictive of treatment selection to minimize confounding factors and showed good prediction performance. The study demonstrated that antidepressant response could be accurately predicted using real-world EHR data with AI modeling, suggesting the potential for developing clinical decision support systems for more effective treatment selection. While considerable progress has been made in leveraging AI techniques and genomics to forecast treatment outcomes, it is essential to conduct further prospective and retrospective clinical research and studies [ 47 , 50 ]. These endeavors are necessary for generating the comprehensive data required to train the algorithms effectively, ensure their reliability in real-world settings, and further develop AI-based clinical decision tools.

Dose optimization and therapeutic drug monitoring

AI plays a crucial role in dose optimization and adverse drug event prediction, offering significant benefits in enhancing patient safety and improving treatment outcomes [ 53 ]. By leveraging AI algorithms, healthcare providers can optimize medication dosages tailored to individual patients and predict potential adverse drug events, thereby reducing risks and improving patient care. In a study that aimed to develop an AI-based prediction model for prothrombin time international normalized ratio (PT/INR) and a decision support system for warfarin maintenance dose optimization [ 54 ] The authors analyzed data from 19,719 inpatients across three institutions, and the algorithm outperformed expert physicians with significant differences in predicting future PT/INRs and the generated individualized warfarin dose was reliable.

On the contrary, a novel dose optimization system—CURATE.AI—is an AI-derived platform for dynamically optimizing chemotherapy doses based on individual patient data [ 55 ]. A study was conducted to validate this system as an open-label, prospective trial in patients with advanced solid tumors treated with three different chemotherapy regimens. CURATE.AI generated personalized doses for subsequent cycles based on the correlation between chemotherapy dose variation and tumor marker readouts. The integration of CURATE.AI into the clinical workflow showed successful incorporation and potential benefits in terms of reducing chemotherapy dose and improving patient response rates and durations compared to the standard of care. These findings support the need for prospective validation through randomized clinical trials and indicate the potential of AI in optimizing chemotherapy dosing and lowering the risk of adverse drug events.

Therapeutic drug monitoring (TDM) is a process used to optimize drug dosing in individual patients. It is predominantly utilized for drugs with a narrow therapeutic index to avoid both underdosing insufficiently medicating as well as toxic levels. TDM aims to ensure that patients receive the right drug, at the right dose, at the right time, to achieve the desired therapeutic outcome while minimizing adverse effects [ 56 ]. The use of AI in TDM has the potential to revolutionize how drugs are monitored and prescribed. AI algorithms can be trained to predict an individual’s response to a given drug based on their genetic makeup, medical history, and other factors. This personalized approach to drug therapy can lead to more effective treatments and better patient outcomes [ 57 , 58 ].

One example of AI in TDM is using ML algorithms to predict drug-drug interactions. By analyzing large datasets of patient data, these algorithms can identify potential drug interactions. This can help to reduce the risk of adverse drug reactions, and cost and improve patient outcomes [ 59 ]. Another application of AI in TDM using predictive analytics to identify patients at high risk of developing adverse drug reactions. By analyzing patient data and identifying potential risk factors, healthcare providers can take proactive steps to prevent adverse events before they occur [ 60 ]. Overall, the use of AI in TDM has the potential to improve patient outcomes, reduce healthcare costs, and enhance the accuracy and efficiency of drug dosing. As this technology continues to evolve, AI will likely play an increasingly important role in the field of TDM.

AI assistance in population health management

Predictive analytics and risk assessment.

Population health management increasingly uses predictive analytics to identify and guide health initiatives. In data analytics, predictive analytics is a discipline that significantly utilizes modeling, data mining, AI, and ML. In order to anticipate the future, it analyzes historical and current data [ 61 , 62 ]. ML algorithms and other technologies are used to analyze data and develop predictive models to improve patient outcomes and reduce costs. One area where predictive analytics can be instrumental is in identifying patients at risk of developing chronic diseases such as endocrine or cardiac diseases. By analyzing data such as medical history, demographics, and lifestyle factors, predictive models can identify patients at higher risk of developing these conditions and target interventions to prevent or treat them [ 61 ]. Predicting hospital readmissions is another area where predictive analytics can be applied. By analyzing patient demographics, medical history, and social health factors, predictive models can identify patients at higher risk of hospital readmissions and target interventions to prevent readmissions (Fig.  3 ) [ 62 , 63 , 64 ]; this can help reduce healthcare costs and improve patient outcomes which is the reason behind launching new companies such as “Reveal ®” [ 65 ].

Unlocking the Power of Patient Data with AI-Driven Predictive Analytics

AI can be used to optimize healthcare by improving the accuracy and efficiency of predictive models. AI algorithms can analyze large amounts of data and identify patterns and relationships that may not be obvious to human analysts; this can help improve the accuracy of predictive models and ensure that patients receive the most appropriate interventions. AI can also automate specific public health management tasks, such as patient outreach and care coordination [ 61 , 62 ]. Which can help reduce healthcare costs and improve patient outcomes by ensuring patients receive timely and appropriate care. However, it is pivotal to note that the success of predictive analytics in public health management depends on the quality of data and the technological infrastructure used to develop and implement predictive models. In addition, human supervision is vital to ensure the appropriateness and effectiveness of interventions for at-risk patients. In summary, predictive analytics plays an increasingly important role in population health. Using ML algorithms and other technologies, healthcare organizations can develop predictive models that identify patients at risk for chronic disease or readmission to the hospital [ 61 , 62 , 63 , 64 ].

Furthermore, AI is needed to address these challenges regarding vaccine production and supply chain bottlenecks. Testing algorithms on real-time vaccine supply chains can be challenging. To overcome this, investing in research and development is essential to create robust algorithms that can accurately predict and optimize vaccine supply chains. Edge analytics can also detect anomalies and predict Disease X events and associated risks to the healthcare system [ 66 ].

From a Saudi perspective, Sehaa, a big data analytics tool in Saudi Arabia, uses Twitter data to detect diseases, and it found that dermal diseases, heart diseases, hypertension, cancer, and diabetes are the top five diseases in the country [ 67 ]. Riyadh has the highest awareness-to-afflicted ratio for six of the fourteen diseases detected, while Taif is the healthiest city with the lowest number of disease cases and a high number of awareness activities. These findings highlight the potential of predictive analytics in population health management and the need for targeted interventions to prevent and treat chronic diseases in Saudi Arabia [ 67 ]. AI can optimize health care by improving the accuracy and efficiency of predictive models and automating certain tasks in population health management [ 62 ]. However, successfully implementing predictive analytics requires high-quality data, advanced technology, and human oversight to ensure appropriate and effective interventions for patients.

Establishment of working groups, guidelines, and frameworks

AI is transforming how guidelines are established in various fields. In healthcare, guidelines usually take much time, from establishing the knowledge gap that needs to be fulfilled to publishing and disseminating these guidelines. AI can help identify newly published data based on data from clinical trials and real-world patient outcomes within the same area of interest which can then facilitate the first stage of mining information. Then, under the supervision of scientists and experts in the field, AI algorithms can analyze vast amounts of data to identify patterns and trends that can inform the development of evidence-based guidelines in real-time, which allows for a fast exchange of information with essential supervision clinicians for its clinical and ethical implications [ 68 , 69 , 70 , 71 , 72 , 73 ].

Several professional organizations have developed frameworks for addressing concerns unique to developing, reporting, and validating AI in medicine [ 69 , 70 , 71 , 72 , 73 ]. Instead of focusing on the clinical application of AI, these frameworks are more concerned with educating the technological creators of AI by providing instructions on encouraging transparency in the design and reporting of AI algorithms [ 69 ]. Additionally, regulatory regulation of AI is still in its infancy. The US Food and Drug Administration (FDA) is now developing guidelines on critically assessing real-world applications of AI in medicine while publishing a framework to guide the role of AI and ML in software as medical devices [ 74 ]. The European Commission has spearheaded a multidisciplinary effort to improve the credibility of AI [ 75 ], and the European Medicines Agency (EMA) has deemed the regulation of AI a strategic priority [ 76 ]. These legislative efforts are meant to shape the healthcare future to be better equipped to be a technology-driven sector. Overall, the role of AI in establishing guidelines is to provide data-driven insights and recommendations based on vast amounts of information, which can lead to more efficient and effective decision-making, better outcomes, and reduced costs. However, it is crucial to ensure that AI-based guidelines are transparent, fair, unbiased, and informed by human expertise and ethical considerations [ 68 ].

AI in drug information and consultation

AI would propose a new support system to assist practical decision-making tools for healthcare providers. In recent years, healthcare institutions have provided a greater leveraging capacity of utilizing automation-enabled technologies to boost workflow effectiveness and reduce costs while promoting patient safety, accuracy, and efficiency [ 77 ]. By introducing advanced technologies like NLP, ML, and data analytics, AI can significantly provide real-time, accurate, and up-to-date information for practitioners at the hospital. According to the McKinsey Global Institute, ML and AI in the pharmaceutical sector have the potential to contribute approximately $100 billion annually to the US healthcare system [ 78 ]. Researchers claim that these technologies enhance decision-making, maximize creativity, increase the effectiveness of research and clinical trials, and produce new tools that benefit healthcare providers, patients, insurers, and regulators [ 78 ]. AI enables quick and comprehensive retrieval of drug-related information from different resources through its ability to analyze the current medical literature, drug databases, and clinical guidelines to provide accurate and evidence-based decisions for healthcare providers. Using automated response systems, AI-powered virtual assistants can handle common questions and provide detailed medical information to healthcare providers [ 79 ]. AI-powered chatbots help reduce the workload on healthcare providers, allowing them to focus on more complicated cases that require their expertise. Also, AI algorithms can generate specific recommendations for individual patients, considering factors like health conditions, past medical and medication history, and social/lifestyle preferences, allowing healthcare professionals to optimize medication choices and dosages [ 80 , 81 ].

AI-powered patient care

Ai virtual healthcare assistance.

With continuously increasing demands of health care services and limited resources worldwide, finding solutions to overcome these challenges is essential [ 82 ]. Virtual health assistants are a new and innovative technology transforming the healthcare industry to support healthcare professionals. It is designed to simulate human conversation to offer personalized patient care based on input from the patient [ 83 ]. These digital assistants use AI-powered applications, chatbots, sounds, and interfaces. Virtual assistants can help patients with tasks such as identifying the underlying problem based on the patient’s symptoms, providing medical advice, reminding patients to take their medications, scheduling doctor appointments, and monitoring vital signs. In addition, digital assistants can collect information daily regarding patients’ health and forward the reports to the assigned physician. By taking off some of these responsibilities from human healthcare providers, virtual assistants can help to reduce their workload and improve patient outcomes.

Furthermore, these tools can always be available, making it easier for patients to access healthcare when needed [ 84 ]. Another medical service that an AI-driven phone application can provide is triaging patients and finding out how urgent their problem is, based on the entered symptoms into the app. The National Health Service (NHS) has tested this app in north London, and now about 1.2 million people are using this AI chatbot to answer their questions instead of calling the NHS non-emergency number [ 85 ]. In addition, introducing intelligent speakers into the market has a significant benefit in the lives of elderly and chronically ill patients who are unable to use smartphone apps efficiently [ 86 ]. Overall, virtual health assistants have the potential to significantly improve the quality, efficiency, and cost of healthcare delivery while also increasing patient engagement and providing a better experience for them.

AI mental health support

AI has the potential to revolutionize mental health support by providing personalized and accessible care to individuals [ 87 , 88 ]. Several studies showed the effectiveness and accessibility of using Web-based or Internet-based cognitive-behavioral therapy (CBT) as a psychotherapeutic intervention [ 89 , 90 ]. Even though psychiatric practitioners rely on direct interaction and behavioral observation of the patient in clinical practice compared to other practitioners, AI-powered tools can supplement their work in several ways. AI-powered mental health applications can assist in the early detection and diagnosis of mental health conditions, as well as provide tailored treatment and support [ 88 , 89 , 90 , 91 ] These applications can also offer round-the-clock support, reducing the need for in-person appointments and wait times. Furthermore, these digital tools can be used to monitor patient progress and medication adherence, providing valuable insights into treatments’ effectiveness [ 88 ].

The current published studies addressing the applicability of AI in mental health concluded that depression is the most commonly investigated mental disorder [ 88 ]. Moreover, AI-powered apps prove their benefits in patients with substance use disorder. A recent study evaluated the utility of a mental health digital app called Woebot in patients with substance use disorders. This study found that using Woebot was significantly associated with improved substance use, cravings, depression, and anxiety [ 92 ]. While AI-powered mental health diagnosis holds promise, some significant limitations must be addressed. One of the main limitations is the risk of bias in the data and algorithms used in AI-powered diagnosis. If the data used to train AI algorithms does not represent diverse populations, it can lead to biased and inaccurate results. Additionally, AI-powered diagnosis may not take into account the complexity of mental health conditions, which can present differently in different people. Finally, there is a risk that AI-powered diagnosis may lead to a lack of personalization and empathy in mental health care, which is an important aspect of successful treatment [ 93 ]. Therefore, while AI-powered diagnosis can be a valuable tool in mental health care, it should be used as a supplement to, rather than a replacement for, professional diagnosis and treatment.

AI in enhancing patient education and mitigating healthcare provider burnout

One of the emerging applications of AI is patient education [ 94 ]. AI-powered chatbots are being implemented in various healthcare contexts, such as diet recommendations [ 95 , 96 ], smoking cessation, and cognitive-behavioral therapy [ 97 ]. Patient education is integral to healthcare, as it enables individuals to understand their medical diagnosis, treatment options, and preventative measures [ 98 ]. Informed patients are more likely to adhere to their treatment regimens and achieve better health outcomes [ 99 ]. AI has the potential to play a significant role in patient education by providing personalized and interactive information and guidance to patients and their caregivers [ 100 ]. For example, in patients with prostate cancer, introducing a prostate cancer communication assistant (PROSCA) chatbot offered a clear to moderate increase in participants’ knowledge about prostate cancer [ 101 ]. Researchers found that ChatGPT, an AI Chatbot founded by OpenAI, can help patients with diabetes understand their diagnosis and treatment options, monitor their symptoms and adherence, provide feedback and encouragement, and answer their questions [ 102 ]. AI technology can also be applied to rewrite patient education materials into different reading levels. This suggests that AI can empower patients to take greater control of their health by ensuring that patients can understand their diagnosis, treatment options, and self-care instructions [ 103 ]. However, there are also some challenges and limitations that need to be addressed, such as ensuring the accuracy, reliability, and transparency of the information provided by AI, respecting the privacy and confidentiality of the patients’ data, and maintaining a human touch and empathy in the communication [ 104 ]. The use of AI in patient education is still in its early stages, but it has the potential to revolutionize the way that patients learn about their health. As AI technology continues to develop, we can expect to see even more innovative and effective ways to use AI to educate patients.

Are individuals more inclined towards AI than human healthcare providers

Public perception of the benefits and risks of AI in healthcare systems is a crucial factor in determining its adoption and integration. People’s feelings about AI replacing or augmenting human healthcare practitioners, its role in educating and empowering patients, and its impact on the quality and efficiency of care, as well as on the well-being of healthcare workers, are all important considerations. In medicine, patients often trust medical staff unconditionally and believe that their illness will be cured due to a medical phenomenon known as the placebo effect. In other words, patient-physician trust is vital in improving patient care and the effectiveness of their treatment [ 105 ]. For the relationship between patients and an AI-based healthcare delivery system to succeed, building a relationship based on trust is imperative [ 106 ].

Research on whether people prefer AI over healthcare practitioners has shown mixed results depending on the context, type of AI system, and participants’ characteristics [ 107 , 108 ]. Some surveys have indicated that people are generally willing to use or interact with AI for health-related purposes such as diagnosis, treatment, monitoring, or decision support [ 108 , 109 , 110 ]. However, other studies have suggested that people still prefer human healthcare practitioners over AI, especially for complex or sensitive issues such as mental health, chronic diseases, or end-of-life care [ 108 , 111 ]. In a US-based study, 60% of participants expressed discomfort with providers relying on AI for their medical care. However, the same study found that 80% of Americans would be willing to use AI-powered tools to help manage their health [ 109 ]. In another survey, responders’ comfort with AI varied based on clinical application, and most patients felt that AI would improve their healthcare, which suggests that people are generally willing to use AI for healthcare-related purposes and that patient education, concerns, and comfort levels should be accounted for when planning for integration of AI [ 110 ]. Moreover, people’s trust and acceptance of AI may vary depending on their age, gender, education level, cultural background, and previous experience with technology [ 111 , 112 ].

Future directions and considerations for clinical implementation

Obstacles and solutions.

AI has the potential to revolutionize clinical practice, but several challenges must be addressed to realize its full potential. Among these challenges is the lack of quality medical data, which can lead to inaccurate outcomes. Data privacy, availability, and security are also potential limitations to applying AI in clinical practice. Additionally, determining relevant clinical metrics and selecting an appropriate methodology is crucial to achieving the desired outcomes. Human contribution to the design and application of AI tools is subject to bias and could be amplified by AI if not closely monitored [ 113 ]. The AI-generated data and/or analysis could be realistic and convincing; however, hallucination could also be a major issue which is the tendency to fabricate and create false information that cannot be supported by existing evidence [ 114 ]. This can be particularly problematic regarding sensitive areas such as patient care. Thus, the development of AI tools has implications for current health professions education, highlighting the necessity of recognizing human fallibility in areas including clinical reasoning and evidence-based medicine [ 115 ]. Finally, human expertise and involvement are essential to ensure the appropriate and practical application of AI to meet clinical needs and the lack of this expertise could be a drawback for the practical application of AI.

Addressing these challenges and providing constructive solutions will require a multidisciplinary approach, innovative data annotation methods, and the development of more rigorous AI techniques and models. Creating practical, usable, and successfully implemented technology would be possible by ensuring appropriate cooperation between computer scientists and healthcare providers. By merging current best practices for ethical inclusivity, software development, implementation science, and human-computer interaction, the AI community will have the opportunity to create an integrated best practice framework for implementation and maintenance [ 116 ]. Additionally, a collaboration between multiple health care settings is required to share data and ensure its quality, as well as verify analyzed outcomes which will be critical to the success of AI in clinical practice. Another suggestion is to provide appropriate training and education that starts from the undergraduate level of all the health care practitioners and proceeds to the continuous development and improvement for the practitioners working in the current practice to ensure the proper adaptation which provides the best patient care and avoid any legal or ethical issues or misinterpretation of the outcomes without verifying the results [ 117 ]. Medical schools are encouraged to incorporate AI-related topics into their medical curricula. A study conducted among radiology residents showed that 86% of students agreed that AI would change and improve their practice, and up to 71% felt that AI should be taught at medical schools for better understanding and application [ 118 ]. This integration ensures that future healthcare professionals receive foundational knowledge about AI and its applications from the early stages of their education.

Legal, ethical, and risk associated with AI in healthcare system

Converting AI and big data into secure and efficient practical applications, services, and procedures in healthcare involves significant costs and risks. Consequently, safeguarding the commercial interests of AI and data-driven healthcare technologies has emerged as an increasingly crucial subject [ 119 ]. In the past, only medical professionals could measure vital signs such as blood pressure, glucose levels, and heart rate [ 48 ]. However, contemporary mobile applications now enable the continuous collection of such information. Nevertheless, addressing the ethical risks associated with AI implementation is imperative, particularly concerning data privacy and confidentiality violations, informed consent, and patient autonomy [ 48 , 119 ]. Given the prominence of big data and AI in healthcare and precision medicine, robust data protection legislation becomes paramount to safeguarding individual privacy. Countries around the world have introduced laws to protect the privacy of their citizens, such as the Health Insurance Portability and Accountability Act (HIPAA) in the United States and the General Data Protection Regulation (GDPR) in Europe [ 120 , 121 ]. While HIPAA protects only relevant health information produced by covered entities, the GDPR has implemented extensive data protection law within the EU, creating a significant global shift in data protection [ 120 , 121 ].

One of the major causes that can compromise patient data, disrupt critical healthcare operations, and jeopardize patient safety with the use of AI in the healthcare system is increased cyberattacks [ 66 , 122 ]. Predictive algorithms can be employed to detect and prevent these cyber threats. To safeguard data privacy and maintain system integrity, it’s essential to deeply investigate cybersecurity and the cyber risk landscape of healthcare systems [ 66 , 123 ]. By implementing a variety of robust AI algorithms, the risk associated with relying on a singular solution can be mitigated [ 66 , 123 ]. While data privacy and security breaches are challenges associated with AI in healthcare [ 122 ], it offers significant advantages such as task streamlining, enhanced efficiency, time and resource savings, research support, and reduced physician stress [ 122 ]. In the context of ethical considerations, an epistemological framework for ethical assessment has been proposed to prioritize ethical awareness, transparency, and accountability when evaluating digital technology’s impact on healthcare supply chain participants [ 123 , 124 ].

The integration of AI in healthcare has immense potential to revolutionize patient care and outcomes. AI-driven predictive analytics can enhance the accuracy, efficiency, and cost-effectiveness of disease diagnosis and clinical laboratory testing. Additionally, AI can aid in population health management and guideline establishment, providing real-time, accurate information and optimizing medication choices. Integrating AI in virtual health and mental health support has shown promise in improving patient care. However, it is important to address limitations such as bias and lack of personalization to ensure equitable and effective use of AI.

Several measures must be taken to ensure responsible and effective implementation of AI in healthcare.

Firstly, comprehensive cybersecurity strategies and robust security measures should be developed and implemented to protect patient data and critical healthcare operations. Collaboration between healthcare organizations, AI researchers, and regulatory bodies is crucial to establishing guidelines and standards for AI algorithms and their use in clinical decision-making. Investment in research and development is also necessary to advance AI technologies tailored to address healthcare challenges.

AI algorithms can continuously examine factors such as population demographics, disease prevalence, and geographical distribution. This can identify patients at a higher risk of certain conditions, aiding in prevention or treatment. Edge analytics can also detect irregularities and predict potential healthcare events, ensuring that resources like vaccines are available where most needed.

Public perception of AI in healthcare varies, with individuals expressing willingness to use AI for health purposes while still preferring human practitioners in complex issues. Trust-building and patient education are crucial for the successful integration of AI in healthcare practice. Overcoming challenges like data quality, privacy, bias, and the need for human expertise is essential for responsible and effective AI integration.

Collaboration among stakeholders is vital for robust AI systems, ethical guidelines, and patient and provider trust. Continued research, innovation, and interdisciplinary collaboration are important to unlock the full potential of AI in healthcare. With successful integration, AI is anticipated to revolutionize healthcare, leading to improved patient outcomes, enhanced efficiency, and better access to personalized treatment and quality care.

Data availability

Not applicable.

Abbreviations

Artificial Intelligence

Autism Spectrum Disorder

Convolutional Neural Networks

Digital in-line holographic microscopy

General Data Protection Regulation

Electronic Health Record

Health Insurance Portability and Accountability Act

Large Language Models

Machine Learning

Natural Language Processing

Prostate cancer communication assistant

Prothrombin Time / International Normalized Ratio

Support vector Machines

Therapeutic drug monitoring

Suleimenov IE, Vitulyova YS, Bakirov AS, Gabrielyan OA. Artificial Intelligence:what is it? Proc 2020 6th Int Conf Comput Technol Appl. 2020;22–5. https://doi.org/10.1145/3397125.3397141 .

Davenport T, Kalakota R. The potential for artificial intelligence in Healthcare. Future Healthc J. 2019;6(2):94–8. https://doi.org/10.7861/futurehosp.6-2-94 .

Article   Google Scholar  

Russell SJ. Artificial intelligence a modern approach. Pearson Education, Inc.; 2010.

McCorduck P, Cfe C. Machines who think: a personal inquiry into the history and prospects of Artificial Intelligence. AK Peters; 2004.

Jordan MI, Mitchell TM. Machine learning: Trends, perspectives, and prospects. Science. 2015;349(6245):255–60. https://doi.org/10.1126/science.aaa8415 .

VanLEHN K. The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems. Educational Psychol. 2011;46(4):197–221. https://doi.org/10.1080/00461520.2011.611369 .

Topol EJ. High-performance medicine: the convergence of human and Artificial Intelligence. Nat Med. 2019;25(1):44–56. https://doi.org/10.1038/s41591-018-0300-7 .

Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017;542(7639):115–8. https://doi.org/10.1038/nature21056 .

Myszczynska MA, Ojamies PN, Lacoste AM, Neil D, Saffari A, Mead R, et al. Applications of machine learning to diagnosis and treatment of neurodegenerative Diseases. Nat Reviews Neurol. 2020;16(8):440–56. https://doi.org/10.1038/s41582-020-0377-8 .

Ahsan MM, Luna SA, Siddique Z. Machine-learning-based disease diagnosis: a comprehensive review. Healthcare. 2022;10(3):541. https://doi.org/10.3390/healthcare10030541 .

McKinney SM, Sieniek M, Godbole V, Godwin J, Antropova N, Ashrafian H, et al. International evaluation of an AI system for breast cancer screening. Nature. 2020;577(7788):89–94. https://doi.org/10.1038/s41586-019-1799-6 .

Kim H-E, Kim HH, Han B-K, Kim KH, Han K, Nam H, et al. Changes in cancer detection and false-positive recall in mammography using Artificial Intelligence: a retrospective, Multireader Study. Lancet Digit Health. 2020;2(3). https://doi.org/10.1016/s2589-7500(20)30003-0 .

Han SS, Park I, Eun Chang S, Lim W, Kim MS, Park GH, et al. Augmented Intelligence Dermatology: deep neural networks Empower Medical Professionals in diagnosing skin Cancer and Predicting Treatment Options for 134 skin Disorders. J Invest Dermatol. 2020;140(9):1753–61. https://doi.org/10.1016/j.jid.2020.01.019 .

Haenssle HA, Fink C, Schneiderbauer R, Toberer F, Buhl T, Blum A, et al. Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists. Ann Oncol. 2018;29(8):1836–42. https://doi.org/10.1093/annonc/mdy166 .

Li S, Zhao R, Zou H. Artificial intelligence for diabetic retinopathy. Chin Med J (Engl). 2021;135(3):253–60. https://doi.org/10.1097/CM9.0000000000001816 .

Alfaras M, Soriano MC, Ortín S. A fast machine learning model for ECG-based Heartbeat classification and arrhythmia detection. Front Phys. 2019;7. https://doi.org/10.3389/fphy.2019.00103 .

Raghunath S, Pfeifer JM, Ulloa-Cerna AE, Nemani A, Carbonati T, Jing L, et al. Deep neural networks can predict new-onset atrial fibrillation from the 12-lead ECG and help identify those at risk of atrial fibrillation–related stroke. Circulation. 2021;143(13):1287–98. https://doi.org/10.1161/circulationaha.120.047829 .

Becker J, Decker JA, Römmele C, Kahn M, Messmann H, Wehler M, et al. Artificial intelligence-based detection of pneumonia in chest radiographs. Diagnostics. 2022;12(6):1465. https://doi.org/10.3390/diagnostics12061465 .

Mijwil MM, Aggarwal K. A diagnostic testing for people with appendicitis using machine learning techniques. Multimed Tools Appl. 2022;81(5):7011–23. https://doi.org/10.1007/s11042-022-11939-8 .

Undru TR, Uday U, Lakshmi JT, et al. Integrating Artificial Intelligence for Clinical and Laboratory diagnosis - a review. Maedica (Bucur). 2022;17(2):420–6. https://doi.org/10.26574/maedica.2022.17.2.420 .

Peiffer-Smadja N, Dellière S, Rodriguez C, Birgand G, Lescure FX, Fourati S, et al. Machine learning in the clinical microbiology laboratory: has the time come for routine practice? Clin Microbiol Infect. 2020;26(10):1300–9. https://doi.org/10.1016/j.cmi.2020.02.006 .

Smith KP, Kang AD, Kirby JE. Automated interpretation of Blood Culture Gram Stains by Use of a deep convolutional neural network. J Clin Microbiol. 2018;56(3):e01521–17. https://doi.org/10.1128/JCM.01521-17 .

Weis CV, Jutzeler CR, Borgwardt K. Machine learning for microbial identification and antimicrobial susceptibility testing on MALDI-TOF mass spectra: a systematic review. Clin Microbiol Infect. 2020;26(10):1310–7. https://doi.org/10.1016/j.cmi.2020.03.014 .

Go T, Kim JH, Byeon H, Lee SJ. Machine learning-based in-line holographic sensing of unstained malaria-infected red blood cells. J Biophotonics. 2018;11(9):e201800101. https://doi.org/10.1002/jbio.201800101 .

Smith KP, Kirby JE. Image analysis and artificial intelligence in infectious disease diagnostics. Clin Microbiol Infect. 2020;26(10):1318–23. https://doi.org/10.1016/j.cmi.2020.03.012 .

Vandenberg O, Durand G, Hallin M, Diefenbach A, Gant V, Murray P, et al. Consolidation of clinical Microbiology Laboratories and introduction of Transformative Technologies. Clin Microbiol Rev. 2020;33(2). https://doi.org/10.1128/cmr.00057-19 .

Panch T, Szolovits P, Atun R. Artificial Intelligence, Machine Learning and Health Systems. J Global Health. 2018;8(2). https://doi.org/10.7189/jogh.08.020303 .

Berlyand Y, Raja AS, Dorner SC, Prabhakar AM, Sonis JD, Gottumukkala RV, et al. How artificial intelligence could transform emergency department operations. Am J Emerg Med. 2018;36(8):1515–7. https://doi.org/10.1016/j.ajem.2018.01.017 .

Matheny ME, Whicher D, Thadaney Israni S. Artificial Intelligence in Health Care: a Report from the National Academy of Medicine. JAMA. 2020;323(6):509–10. https://doi.org/10.1001/jama.2019.21579 .

Jiang F, Jiang Y, Zhi H, Dong Y, Li H, Ma S, et al. Artificial intelligence in healthcare: past, present and future. Stroke Vasc Neurol. 2017;2(4):230–43. https://doi.org/10.1136/svn-2017-000101 .

Gandhi SO, Sabik L. Emergency department visit classification using the NYU algorithm. Am J Manag Care. 2014;20(4):315–20.

Google Scholar  

Hautz WE, Kämmer JE, Hautz SC, Sauter TC, Zwaan L, Exadaktylos AK, et al. Diagnostic error increases mortality and length of hospital stay in patients presenting through the emergency room. Scand J Trauma Resusc Emerg Med. 2019;27(1):54. https://doi.org/10.1186/s13049-019-0629-z .

Haug CJ, Drazen JM. Artificial Intelligence and Machine Learning in Clinical Medicine, 2023. N Engl J Med. 2023;388(13):1201–8. https://doi.org/10.1056/NEJMra2302038 .

Abubaker Bagabir S, Ibrahim NK, Abubaker Bagabir H, Hashem Ateeq R. Covid-19 and Artificial Intelligence: genome sequencing, drug development and vaccine discovery. J Infect Public Health. 2022;15(2):289–96. https://doi.org/10.1016/j.jiph.2022.01.011 .

Pudjihartono N, Fadason T, Kempa-Liehr AW, O’Sullivan JM. A review of feature selection methods for machine learning-based Disease Risk Prediction. Front Bioinform. 2022;2:927312. https://doi.org/10.3389/fbinf.2022.927312 . Published 2022 Jun 27.

Widen E, Raben TG, Lello L, Hsu SDH. Machine learning prediction of biomarkers from SNPs and of Disease risk from biomarkers in the UK Biobank. Genes (Basel). 2021;12(7):991. https://doi.org/10.3390/genes12070991 . Published 2021 Jun 29.

Wang H, Avillach P. Diagnostic classification and prognostic prediction using common genetic variants in autism spectrum disorder: genotype-based Deep Learning. JMIR Med Inf. 2021;9(4). https://doi.org/10.2196/24754 .

Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci. 2001;98:10869–74. https://doi.org/10.1073/pnas.191367098 .

Yersal O. Biological subtypes of breast cancer: prognostic and therapeutic implications. World J Clin Oncol. 2014;5(3):412–24. https://doi.org/10.5306/wjco.v5.i3.412 .

eek JT, Scharpf RB, Bravo HC, Simcha D, Langmead B, Johnson WE, et al. Tackling the widespread and critical impact of batch effects in high-throughput data. Nat Rev Genet. 2010;11:733–9. https://doi.org/10.1038/nrg2825 .

Blanco-González A, Cabezón A, Seco-González A, Conde-Torres D, Antelo-Riveiro P, Piñeiro Á, et al. The role of AI in drug discovery: Challenges, opportunities, and strategies. Pharmaceuticals. 2023;16(6):891. https://doi.org/10.3390/ph16060891 .

Tran TTV, Surya Wibowo A, Tayara H, Chong KT. Artificial Intelligence in Drug Toxicity Prediction: recent advances, Challenges, and future perspectives. J Chem Inf Model. 2023;63(9):2628–43. https://doi.org/10.1021/acs.jcim.3c00200 .

Tran TTV, Tayara H, Chong KT. Artificial Intelligence in Drug Metabolism and Excretion Prediction: recent advances, Challenges, and future perspectives. Pharmaceutics. 2023;15(4):1260. https://doi.org/10.3390/pharmaceutics15041260 .

Guedj M, Swindle J, Hamon A, Hubert S, Desvaux E, Laplume J, et al. Industrializing AI-powered drug discovery: Lessons learned from the patrimony computing platform. Expert Opin Drug Discov. 2022;17(8):815–24. https://doi.org/10.1080/17460441.2022.2095368 .

Ahmed F, Kang IS, Kim KH, Asif A, Rahim CS, Samantasinghar A, et al. Drug repurposing for viral cancers: a paradigm of machine learning, Deep Learning, and virtual screening-based approaches. J Med Virol. 2023;95(4). https://doi.org/10.1002/jmv.28693 .

Singh DP, Kaushik B. A systematic literature review for the prediction of anticancer drug response using various machine-learning and deep-learning techniques. Chem Biol Drug Des. 2023;101(1):175–94. https://doi.org/10.1111/cbdd.14164 .

Quazi S. Artificial intelligence and machine learning in precision and genomic medicine. Med Oncol. 2022;39(2):120. https://doi.org/10.1007/s12032-022-01711-1 .

Subramanian M, Wojtusciszyn A, Favre L, Boughorbel S, Shan J, Letaief KB, et al. Precision medicine in the era of artificial intelligence: implications in chronic disease management. J Transl Med. 2020;18(1):472. https://doi.org/10.1186/s12967-020-02658-5 .

Johnson KB, Wei WQ, Weeraratne D, Frisse ME, Misulis K, Rhee K, et al. Precision Medicine, AI, and the future of Personalized Health Care. Clin Transl Sci. 2021;14(1):86–93. https://doi.org/10.1111/cts.12884 .

Pulley JM, Denny JC, Peterson JF, Bernard GR, Vnencak-Jones CL, Ramirez AH, et al. Operational implementation of prospective genotyping for personalized medicine: the design of the Vanderbilt PREDICT project. Clin Pharmacol Ther. 2012;92(1):87–95. https://doi.org/10.1038/clpt.2011.371 .

Huang C, Clayton EA, Matyunina LV, McDonald LD, Benigno BB, Vannberg F, et al. Machine learning predicts individual cancer patient responses to therapeutic drugs with high accuracy. Sci Rep. 2018;8(1):16444. https://doi.org/10.1038/s41598-018-34753-5 .

Sheu YH, Magdamo C, Miller M, Das S, Blacker D, Smoller JW. AI-assisted prediction of differential response to antidepressant classes using electronic health records. npj Digit Med. 2023;6:73. https://doi.org/10.1038/s41746-023-00817-8 .

Martin GL, Jouganous J, Savidan R, Bellec A, Goehrs C, Benkebil M, et al. Validation of Artificial Intelligence to support the automatic coding of patient adverse drug reaction reports, using Nationwide Pharmacovigilance Data. Drug Saf. 2022;45(5):535–48. https://doi.org/10.1007/s40264-022-01153-8 .

Lee H, Kim HJ, Chang HW, Kim DJ, Mo J, Kim JE. Development of a system to support warfarin dose decisions using deep neural networks. Sci Rep. 2021;11(1):14745. Published 2021 Jul 20. https://doi.org/10.1038/s41598-021-94305-2 .

Blasiak A, Truong A, Jeit W, Tan L, Kumar KS, Tan SB, et al. PRECISE CURATE.AI: a prospective feasibility trial to dynamically modulate personalized chemotherapy dose with artificial intelligence. J Clin Oncol. 2022;40(16suppl):1574–4. https://doi.org/10.1200/JCO.2022.40.16_suppl.1574 .

Sjövall F, Lanckohr C, Bracht H. What’s new in therapeutic drug monitoring of antimicrobials? Intensive care Med. 2023 May 3. https://doi.org/10.1007/s00134-023-07060-5 .

Partin A, Brettin TS, Zhu Y, Narykov O, Clyde A, Overbeek J, Stevens RL. Deep learning methods for drug response prediction in cancer: predominant and emerging trends. Front Med (Lausanne). 2023;10:1086097. https://doi.org/10.3389/fmed.2023.1086097 .

Zhang H, Chen Y, Li F. Predicting Anticancer Drug Response with Deep Learning constrained by signaling pathways. Front Bioinform. 2021;1:639349. https://doi.org/10.3389/fbinf.2021.639349 .

Han K, Cao P, Wang Y, Xie F, Ma J, Yu M, Wang J, Xu Y, Zhang Y, Wan J. A review of approaches for Predicting Drug-Drug interactions based on machine learning. Front Pharmacol. 2022;12:814858. https://doi.org/10.3389/fphar.2021.814858 .

Liu JYH, Rudd JA. Predicting drug adverse effects using a new Gastro-Intestinal Pacemaker Activity Drug Database (GIPADD). Sci Rep. 2023;13(1):6935. https://doi.org/10.1038/s41598-023-33655-5 .

Nelson KM, Chang ET, Zulman DM, Rubenstein LV, Kirkland FD, Fihn SD. Using Predictive Analytics to Guide Patient Care and Research in a National Health System. J Gen Intern Med. 2019;34(8):1379–80. https://doi.org/10.1007/s11606-019-04961-4 .

Amarasingham R, Patzer RE, Huesch M, Nguyen NQ, Xie B. Implementing electronic health care predictive analytics: considerations and challenges. Health Aff (Millwood). 2014;33(7):1148–54. https://doi.org/10.1377/hlthaff.2014.0352 .

Ansari MS, Alok AK, Jain D, et al. Predictive model based on Health Data Analysis for Risk of Readmission in Disease-Specific cohorts. Perspect Health Inf Manag. 2021;18(Spring):1j. Published 2021 Mar 15.

Donzé J, Aujesky D, Williams D, Schnipper JL. Potentially avoidable 30-day hospital readmissions in medical patients: derivation and validation of a prediction model. JAMA Intern Med. 2013;173:632–8.

Predictive Analytics in Healthcare | Reveal. https://www.revealbi.io/blog/predictive-analytics-in-healthcare . Accessed 6.20.2023.

Alotaibi S, Mehmood R, Katib I, Rana O, Albeshri A, Sehaa. A Big Data Analytics Tool for Healthcare symptoms and Diseases Detection using Twitter, Apache Spark, and machine learning. Appl Sci. 2020;10:1398. https://doi.org/10.3390/app10041398 .

Crossnohere NL, Elsaid M, Paskett J, Bose-Brill S, Bridges JFP. Guidelines for Artificial Intelligence in Medicine: Literature Review and Content Analysis of Frameworks. J Med Internet Res. 2022;24(8):e36823. https://doi.org/10.2196/36823 .

Rivera SC, Liu X, Chan A, Denniston AK, Calvert MJ, SPIRIT-AICONSORT-AI. Working Group Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension. BMJ. 2020;370:m3210. https://doi.org/10.1136/bmj.m3210 .

Beam YuK, Kohane AL. Artificial intelligence in healthcare. Nat Biomed Eng. 2018;2(10):719–31. https://doi.org/10.1038/s41551-018-0305-z .

Vollmer S, Mateen BA, Bohner G, Király FJ, Ghani R, Jonsson P, et al. Machine learning and artificial intelligence research for patient benefit: 20 critical questions on transparency, replicability, ethics, and effectiveness. BMJ. 2020;368:l6927. https://doi.org/10.1136/bmj.l6927 .

Collins GS, Dhiman P, Andaur Navarro CL, Ma J, Hooft L, Reitsma JB, et al. Protocol for development of a reporting guideline (TRIPOD-AI) and risk of bias tool (PROBAST-AI) for diagnostic and prognostic prediction model studies based on artificial intelligence. BMJ Open. 2021;11(7):e048008. https://doi.org/10.1136/bmjopen-2020-048008 .

Liu Y, Chen PC, Krause J, Peng L. How to read articles that use machine learning: users’ guides to the medical literature. JAMA. 2019;322(18):1806–16. https://doi.org/10.1001/jama.2019.16489.2754798 .

Artificial Intelligence and Machine Learning in Software as a Medical Device. US Food and Drug Administration. Released 2021. FDA website: https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device . Accessed 6.20.2023.

White Paper on Artificial Intelligence, European Commission. A European approach to excellence and trust. 2020. Feb 19, https://ec.europa.eu/info/publications/white-paper-artificial-intelligence-european-approach-excellence-and-trust_en . Accessed 6.20.2023.

Radanliev P, De Roure D. Disease X vaccine production and supply chains: risk assessing healthcare systems operating with artificial intelligence and industry 4.0. Health Technol (Berl). 2023;13(1):11–5. https://doi.org/10.1007/s12553-022-00722-2 .

Regulatory Science Strategy to 2025. European Medicines Agency. Released 2020. https://www.ema.europa.eu/en/about-us/how-we-work/regulatory-science-strategy#regulatory-science-strategy-to-2025-section . Accessed 6.20.2023.

Dasta J. Application of artificial intelligence to pharmacy and medicine. Hosp Pharm. 1992;27(4):312–5.

Pharma News Intelligence. Available from: https://pharmanewsintel.com/ . Accessed 6.20.2023.

Chatbots. Medicine Delivery.; Available from: https://hellotars.com/chatbot-templates/healthcare/Hk8N4h/medicine-ordering-chatbot . Accessed 6.20.2023.

Li LR, Du B, Liu HQ, Chen C. Artificial Intelligence for Personalized Medicine in thyroid Cancer: current status and future perspectives. Front Oncol. 2021;10:604051. https://doi.org/10.3389/fonc.2020.604051 .

Davoudi A, Malhotra KR, Shickel B, Siegel S, Williams S, Ruppert M et al. The intelligent ICU pilot study: using artificial intelligence technology for autonomous patient monitoring. https://doi.org/10.48550/arXiv.1804.10201 .

Buch VH, Ahmed I, Maruthappu M. Artificial intelligence in medicine: current trends and future possibilities. Br J Gen Pract. 2018;68(668):143–4. https://doi.org/10.3399/bjgp18X695213 .

Curtis RG, Bartel B, Ferguson T, Blake HT, Northcott C, Virgara R, et al. Improving user experience of virtual Health Assistants: scoping review. J Med Internet Res. 2021;23(12):e31737. https://doi.org/10.2196/31737 .

Ghosh PK, Jain P, Wankhede S, Preethi M, Kannan MK. Virtual nursing Assistant. J Geog Sci. 2021;8:279–85. 20.18001.GSJ.2021.V8I3.21.36690.

Burgess M. The NHS is trialling an AI chatbot to answer your medical questions. Wired. 2017. Jan 5, http://www.wired.co.uk/article/babylon-nhs-chatbot-app . Accessed 20 June 2023.

Pavel Jiřík. Inspiring Applications of Digital Virtual Assistants in Healthcare. July 22., 2022. https://www.phonexia.com/blog/inspiring-applications-of-digital-virtual-assistants-in-healthcare/ . Accessed 20 June 2023.

Kim JW, Jones KL, D’Angelo E. How to prepare prospective psychiatrists in the era of Artificial Intelligence. Acad Psychiatry. 2019;43(3):337–9. https://doi.org/10.1007/s40596-019-01025-x .

Graham S, Depp C, Lee EE, Nebeker C, Tu X, Kim HC, et al. Artificial Intelligence for Mental Health and Mental Illnesses: an overview. Curr Psychiatry Rep. 2019;21(11):116. https://doi.org/10.1007/s11920-019-1094-0 .

Fitzpatrick KK, Darcy A, Vierhile M. Delivering cognitive behavior therapy to young adults with symptoms of depression and anxiety using a fully automated conversational agent (woebot): a randomized controlled trial. JMIR Mental Health. 2017;4(2):e19.

Williams AD, Andrews G. The effectiveness of internet cognitive behavioural therapy (iCBT) for depression in primary care: a quality assurance study. PLoS ONE. 2013;8(2):e57447.

Luxton DD. Artificial intelligence in psychological practice: current and future applications and implications. Prof Psychol Res Pract. 2014;45(5):332–9. https://doi.org/10.1037/a0034559 .

Prochaska J, Vogel E, Chieng A, Kendra M, Baiocchi M, Pajarito S, Robinson A. A therapeutic Relational Agent for reducing problematic substance use (woebot): Development and Usability Study. J Med Internet Res. 2021;23(3):e24850.

Lee EE, Torous J, De Choudhury M, Depp CA, Graham SA, Kim HC, et al. Artificial Intelligence for Mental Health Care: clinical applications, barriers, facilitators, and Artificial Wisdom. Biol Psychiatry Cogn Neurosci Neuroimaging. 2021;6(9):856–64. https://doi.org/10.1016/j.bpsc.2021.02.001 .

Artificial Intelligence in Healthcare. 39 Examples Improving the Future of Medicine. Emerj. https://emerj.com/ai-sector-overviews/artificial-intelligence-in-healthcare-39-examples-improving-the-future-of-medicine/ . Published September 21, 2021. Accessed June 19, 2023.

Chew HSJ. The Use of Artificial Intelligence-Based conversational agents (Chatbots) for weight loss: scoping review and practical recommendations. JMIR Med Inform. 2022;10(4):e32578. https://doi.org/10.2196/32578 .

Zhang J, Oh YJ, Lange P, Yu Z, Fukuoka Y. Artificial Intelligence Chatbot Behavior Change Model for Designing Artificial Intelligence Chatbots to promote physical activity and a healthy Diet: viewpoint. J Med Internet Res. 2020;22(9):e22845. https://doi.org/10.2196/22845 .

Wang H, Zhang Z, Ip M, Lau J. T.F. Social media–based conversational agents for health management and interventions. J Med Internet Res. 2018;20(8):e261. https://doi.org/10.2196/jmir.9275 .

Bombard Y, Baker GR, Orlando E, Fancott C, Bhatia P, Casalino S, et al. Engaging patients to improve quality of care: a systematic review. Implement Sci. 2018;13(1):98. https://doi.org/10.1186/s13012-018-0784-z .

Wong CK, Yeung DY, Ho HC, Tse KP, Lam CY. Chinese older adults’ internet use for health information. J Appl Gerontol. 2014;33(3):316–35. https://doi.org/10.1177/0733464812463430 .

Aggarwal A, Tam CC, Wu D, Li X, Qiao S. Artificial Intelligence-Based chatbots for promoting health behavioral changes: systematic review. J Med Internet Res. 2023;25:e40789. https://doi.org/10.2196/40789 .

Görtz M, Baumgärtner K, Schmid T, Muschko M, Woessner P, Gerlach A, et al. An artificial intelligence-based chatbot for prostate cancer education: design and patient evaluation study. Digit Health. 2023;9:20552076231173304. https://doi.org/10.1177/20552076231173304 .

Nakhleh A, Spitzer S, Shehadeh N. ChatGPT’s response to the diabetes knowledge questionnaire: implications for Diabetes Education. Diabetes Technol Ther. 2023 Apr;16. https://doi.org/10.1089/dia.2023.0134 .

irchner GJ, Kim RY, Weddle JB, Bible JE. Can Artificial Intelligence improve the readability of Patient Education Materials? Clin Orthop Relat Res 2023 Apr 28. https://doi.org/10.1097/CORR.0000000000002668 .

Lee D, Yoon SN. Application of Artificial Intelligence-Based Technologies in the Healthcare Industry: Opportunities and Challenges. Int J Environ Res Public Health. 2021;18(1):271. https://doi.org/10.3390/ijerph18010271 .

Kaptchuk TJ, Miller FG. Placebo Effects in Medicine. N Engl J Med. 2015;373(1):8–9. https://doi.org/10.1056/NEJMp1504023 .

Lupton M. Some ethical and legal consequences of the application of artificial intelligence in the field of medicine. Trends in Medicine. 2018;18(4). https://doi.org/10.15761/tim.1000147 .

Pezzo MV, Beckstead JW. Patients prefer artificial intelligence to a human provider, provided the AI is better than the human: A commentary on Longoni, Bonezzi and Morewedge (2019). Judgment and Decision Making. Cambridge University Press; 2020;15(3):443–5. https://doi.org/10.1017/S1930297500007221 .

How Americans View Use of AI in Health Care and Medical Research. https://www.pewresearch.org/science/2023/02/22/60-of-americans-would-be-uncomfortable-with-provider-relying-on-ai-in-their-own-health-care/ . Accessed 19 June 2023.

Esmaeilzadeh P. Use of AI-based tools for healthcare purposes: a survey study from consumers’ perspectives. BMC Med Inform Decis Mak. 2020;20(1):170. Published 2020 Jul 22. https://doi.org/10.1186/s12911-020-01191-1 .

Khullar D, Casalino LP, Qian Y, Lu Y, Krumholz HM, Aneja S. Perspectives of patients about Artificial Intelligence in Health Care. JAMA Netw Open. 2022;5(5):e2210309. https://doi.org/10.1001/jamanetworkopen.2022.10309 . Published 2022 May 2.

Russo S, Jongerius C, Faccio F, et al. Understanding patients’ preferences: a systematic review of Psychological Instruments used in patients’ preference and decision studies. Value Health. 2019;22(4):491–501. https://doi.org/10.1016/j.jval.2018.12.007 .

Young AT, Amara D, Bhattacharya A, Wei ML. Patient and general public attitudes towards clinical artificial intelligence: a mixed methods systematic review. Lancet Digit Health. 2021;3(9):e599–e611. https://doi.org/10.1016/S2589-7500(21)00132-1 .

West SM, Whittaker M, Crawford K. Discriminating Systems: gender, race and power in AI. AI Now Institute; 2019.

Maynez J, Narayan S, Bohnet B, McDonald R. On faithfulness and factuality in abstractive summarization. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020; https://doi.org/10.18653/v1/2020.acl-main.173 .

deBurca S. The learning health care organization. Int J Qual Health Care. 2000;12(6):457–8. https://doi.org/10.1093/intqhc/12.6.457 .

IOM (Institute of Medicine). Measuring the impact of Interprofessional Education on collaborative practice and patient outcomes. Washington, DC: The National Academies Press; 2015. p. 182.

Alqahtani T, Badreldin HA, Alrashed M, Alshaya AI, Alghamdi SS, bin Saleh K, et al. The emergent role of Artificial Intelligence, natural learning processing, and large language models in higher education and research. Res Social Administrative Pharm. 2023. https://doi.org/10.1016/j.sapharm.2023.05.016 .

Pinto dos Santos D, Giese D, Brodehl S, Chon SH, Staab W, Kleinert R, et al. Medical students’ attitude towards Artificial Intelligence: a multicentre survey. Eur Radiol. 2018;29(4):1640–6. https://doi.org/10.1007/s00330-018-5601-1 .

Gerke S, Minssen T, Cohen G. Ethical and legal challenges of artificial intelligence-driven healthcare. Artif Intell Healthc. 2020;295–336. https://doi.org/10.1016/b978-0-12-818438-7.00012-5 .

Cohen IG, Mello MM. HIPAA and protecting health information in the 21st Century. JAMA. 2018;320(3):231. https://doi.org/10.1001/jama.2018.5630 .

Yuan B, Li J. The policy effect of the General Data Protection Regulation (GDPR) on the digital public health sector in the European Union: an empirical investigation. Int J Environ Res Public Health. 2019;16(6):1070. https://doi.org/10.3390/ijerph16061070 .

Abdel-Hameed Al-Mistarehi M, Mijwil M. ; Youssef Filali; Mariem Bounabi; Guma Ali; Mostafa Abotaleb. Artificial Intelligence Solutions for Health 4.0: Overcoming Challenges and Surveying Applications. MJAIH 2023, 2023, 15–20.

Radanliev P, De Roure D. Epistemological and bibliometric analysis of Ethics and Shared responsibility—health policy and IoT Systems. Sustainability. 2021;13(15):8355.

Radanliev P, De Roure D, Ani U, Carvalho G. The ethics of shared Covid-19 risks: an epistemological framework for ethical health technology assessment of risk in vaccine supply chain infrastructures. Health Technol (Berl). 2021;11(5):1083–91.

Download references

Acknowledgements

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author information

Authors and affiliations.

Department of Pharmacy Practice, College of Pharmacy, King Saud bin Abdulaziz University for Health Sciences, Prince Mutib Ibn Abdullah Ibn Abdulaziz Rd, Riyadh, 14611, Saudi Arabia

Shuroug A. Alowais, Nada Alsuhebany, Abdulrahman I. Alshaya, Sumaya N. Almohareb, Atheer Aldairem, Mohammed Alrashed, Khalid Bin Saleh, Hisham A. Badreldin, Majed S. Al Yami, Shmeylan Al Harbi & Abdulkareem M. Albekairy

King Abdullah International Medical Research Center, Riyadh, Saudi Arabia

Shuroug A. Alowais, Sahar S. Alghamdi, Nada Alsuhebany, Tariq Alqahtani, Abdulrahman I. Alshaya, Sumaya N. Almohareb, Atheer Aldairem, Mohammed Alrashed, Khalid Bin Saleh, Hisham A. Badreldin, Majed S. Al Yami, Shmeylan Al Harbi & Abdulkareem M. Albekairy

Pharmaceutical Care Department, King Abdulaziz Medical City, National Guard Health Affairs, Riyadh, Saudi Arabia

Department of Pharmaceutical Sciences, College of Pharmacy, King Saud bin Abdulaziz University for Health Sciences, Riyadh, Saudi Arabia

Sahar S. Alghamdi & Tariq Alqahtani

You can also search for this author in PubMed   Google Scholar

Contributions

Conceptualization: Shuroug A. Alowais; Sahar S. Alghamdi; Hisham A. Badreldin; Tariq Alqahtani.Writing - original draft: Shuroug A. Alowais, Sahar S. Alghamdi, Nada Alsuhebany, Tariq Alqahtani, Abdulrahman I. Alshaya, Sumaya N. Almohareb, Atheer Aldairem, Mohammed Alrashed, Khalid bin Saleh.Writing - review & editing: Shuroug A. Alowais, Sahar S. Alghamdi, Nada Alsuhebany, Tariq Alqahtani, Abdulrahman I. Alshaya, Sumaya N. Almohareb, Atheer Aldairem, Mohammed Alrashed, Khalid bin Saleh, Hisham A. Badreldin, Majed S. Al Yami, Shmeylan Al Harbi, Abdulkareem M. Albekairy.All authors reviewed the manuscript. During the preparation of this work the author(s) used [Chatsonic/ GPT4] to [edit and change format of the paper]. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Corresponding author

Correspondence to Shuroug A. Alowais .

Ethics declarations

Ethics approval and consent to participate, consent for publication, competing interests.

The authors declare that they have no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Cite this article.

Alowais, S.A., Alghamdi, S.S., Alsuhebany, N. et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med Educ 23 , 689 (2023). https://doi.org/10.1186/s12909-023-04698-z

Download citation

Received : 23 June 2023

Accepted : 19 September 2023

Published : 22 September 2023

DOI : https://doi.org/10.1186/s12909-023-04698-z

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Patient care
• Quality of life
• Decision-making
• Personalized treatment plans

BMC Medical Education

ISSN: 1472-6920

• Submission enquiries: [email protected]
• General enquiries: [email protected]

• Conference proceedings
• © 2023

Artificial Intelligence and Sustainable Computing

Proceedings of ICSISCET 2022

• Manjaree Pandit 0 ,
• M. K. Gaur 1 ,
• Sandeep Kumar   ORCID: https://orcid.org/0000-0003-4125-4165 2

Madhav Institute of Technology and Science, Gwalior, India

You can also search for this editor in PubMed   Google Scholar

CHRIST (Deemed to be University), Kengeri Campus, Bengaluru, India

Presents research works in the field of artificial intelligence and sustainable computing

Provides original works presented at ICSISCET 2022 held in Gwalior, India

Serves as a reference for researchers and practitioners in academia and industry

Part of the book series: Algorithms for Intelligent Systems (AIS)

Conference series link(s): ICSISCET: International Conference on Sustainable and Innovative Solutions for Current Challenges in Engineering & Technology

311 Accesses

Conference proceedings info: ICSISCET 2022.

• Table of contents
• Other volumes

About this book

Editors and affiliations, about the editors, bibliographic information, buying options.

• Available as EPUB and PDF
• Read on any device
• Instant download
• Own it forever
• Durable hardcover edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Other ways to access

This is a preview of subscription content, access via your institution .

Table of contents (59 papers)

Front matter, features extraction and analysis of electro myogram signals using time, frequency, and wavelet transform methods.

• Chillakuru Prasad, I. Kullayamma

Lycopersicon Crop Leaf Disease Identification Using Deep Learning

• Barkha M. Joshi, Hetal Bhavsar

Performance Analysis of Satellite Image Classification Using Deep Learning Neural Network

• CH Hussaian Basha, J. Prajakta, V. Prashanth, Shaik Rafikiran, V. S. Patil, B. Srinivasa Varma et al.

Text Visualization of Entire Corpus Through Single Document Input Tools

• Gowri R. Choudhary, Iti Sharma

Facts Devices with Distinctive Positioning for Voltage Regulation

• Hemant Gupta, Yogendra Kumar

Impact of Online Classes Using Machine Learning Algorithms: Estimation, Classification, and Prediction

• A. Ranichitra, A. Mercy Rani

Computational Methods in Computational Fluid Dynamics

• Fouzia Adjailia, Michal Takáč

Analysis of Humidity Effect on Sensitivity of MEMS Cantilever Sensor

• B. Guruprasad, M. S. Shwetha, Arjun Sunil Rao, D. V. Manjunath

Transfer Learning Approach to Detect and Predict the Malaria from Blood Cell Images

• Priyanka Jangde, Manoj Ramaiya

Voting System Based on Blockchain Technology Smart-Contracts

• Ashwani Kumar Pandey, Abhinandan Tripathi, Ashok Kumar Yadav, Shreya Yadav

“Rejuvenation with Modernization of Hydrogenator”; Prediction and Control of Electromagnetic Vibration with Artificial, Development of Stator Frame Design, Artificial Intelligence—A Review Study

• Yejvander Thakur, Geetesh Goga

A Deep Learning Based Neural Network for Detection of Epileptic Seizure

• Hemant Choubey, Alpana Pandey, Vikas Mahor, Rahul Dubey, Amit Kumar Manjhvar, Sushmita Chaudhari

Recent Developments in Generative Adversarial Networks

• Nakul Singh, Sandeep Kumar Parashar

Artificial Intelligence and Machine Learning for Climate Change Mitigation and Adaptation

• Garima Natani

A Model to Identify the Impairment Caused by Smoking to the Oral Cavity

• Gayatri Meghana Gangipamula, Reetu Jain, Syed Abou Iltaf Hussain

An Efficient Approach to Estimate Software Cost by Analogy Using ACO

• L. Karthika, S. Gunasundari

Recent Developments in the Application of Deep Learning to Stock Market Prediction

• Shraddha Jain Sharma, Ratnalata Gupta

Power Quality Enhancement for Grid Connected PV and Wind System with TS Fuzzy Converters

• Vinay Kumar Tatikayala, Shishir Dixit

Improving Right Ventricle Contouring in Cardiac MR Images Using Integrated Approach for Small Datasets

• Anjali Abhijit Yadav, Sanjay Ramchandra Ganorkar

Other Volumes

• Sustainable Computing and Information Technology
• Computational Intelligence
• Machine learning
• Embedded Systems
• VLSI Design
• Intelligent Computing
• Sustainable Engineering Systems
• ICSISCET 2022 Proceedings

Manjaree Pandit, M. K. Gaur

Sandeep Kumar

Manjaree Pandit obtained her M.Tech. degree in Electrical Engineering from Maulana Azad College of Technology, Bhopal, (India) in 1989 and Ph.D. degree from Jiwaji University, Gwalior, (India) in 2001. She is currently working as a dean academics & a professor in Department of Electrical Engineering, M.I.T.S., Gwalior, (India). She is a senior member of IEEE, a recognized reviewer of a few IEEE Transactions, Springer and Elsevier journals, and has published more than 110 papers in reputed international publications. She has been recognized as an outstanding reviewer by Elsevier and received Top Peer Reviewer Award from the international reviewer database site Publons (Clarivate Analytics) for being in top 1% of reviewers in Engineering for year 2017-19 & year 2018-2019 and top Peer Reviewer Award by Publons for being in top 1% of reviewers in cross field in year 2018-19. Her areas of interest are hybrid renewable energy resources integration with the power grid, nature-inspired algorithms, and artificial intelligence applications to electrical power system. She has successfully completed research projects funded from AICTE, DST, and UGC and has published 75 papers in international journals of repute, guided more than 70 PG dissertations and 06 Ph.D. candidates. Her papers are very well cited in Google Scholar, Scopus, and Web of Science. Dr. Manjaree Pandit has also received ISTE, India, National Research Award for year 2002 and 2004, certificate of Merit from Institution of Engineers, India, and UGC Research Award for carrying out postdoctoral work during 2009-2011. She has been involved in conducting a number of national/international conferences, workshops, and short-term courses.

Dr. Sandeep Kumar is currently an associate professor at CHRIST (Deemed to be University), Bangalore. Before joining CHRIST, he worked with ACEIT, Jaipur, Jagannath University Jaipur, and Amity University, Rajasthan. He is an associate editor for Springer's Human-centric Computing and Information Sciences (HCIS) journal. He has published more than seventy research papers in various international journals/conferences and participated in many national and international conferences and workshops. He has authored/edited six books in the field of nature-inspired algorithms, swarm intelligence, soft computing, and computational intelligence.

Book Title : Artificial Intelligence and Sustainable Computing

Book Subtitle : Proceedings of ICSISCET 2022

Editors : Manjaree Pandit, M. K. Gaur, Sandeep Kumar

Series Title : Algorithms for Intelligent Systems

DOI : https://doi.org/10.1007/978-981-99-1431-9

Publisher : Springer Singapore

eBook Packages : Intelligent Technologies and Robotics , Intelligent Technologies and Robotics (R0)

Copyright Information : The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023

Hardcover ISBN : 978-981-99-1430-2 Published: 24 September 2023

Softcover ISBN : 978-981-99-1433-3 Due: 08 October 2024

eBook ISBN : 978-981-99-1431-9 Published: 23 September 2023

Series ISSN : 2524-7565

Series E-ISSN : 2524-7573

Edition Number : 1

Number of Pages : XIV, 753

Number of Illustrations : 94 b/w illustrations, 328 illustrations in colour

Topics : Computational Intelligence , Circuits and Systems , Cyber-physical systems, IoT , Professional Computing , Machine Learning

• Find a journal
• Publish with us

Suggestions or feedback?

MIT News | Massachusetts Institute of Technology

• Machine learning
• Social justice
• Black holes
• Classes and programs

Departments

• Aeronautics and Astronautics
• Brain and Cognitive Sciences
• Architecture
• Political Science
• Mechanical Engineering

Centers, Labs, & Programs

• Abdul Latif Jameel Poverty Action Lab (J-PAL)
• Picower Institute for Learning and Memory
• Lincoln Laboratory
• School of Architecture + Planning
• School of Engineering
• School of Humanities, Arts, and Social Sciences
• Sloan School of Management
• School of Science
• MIT Schwarzman College of Computing

Multi-AI collaboration helps reasoning and factual accuracy in large language models

Press contact :.

Previous image Next image

An age-old adage, often introduced to us during our formative years, is designed to nudge us beyond our self-centered, nascent minds: "Two heads are better than one." This proverb encourages collaborative thinking and highlights the potency of shared intellect.

Fast forward to 2023, and we find that this wisdom holds true even in the realm of artificial intelligence: Multiple language models, working in harmony, are better than one. 

Recently, a team from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) embodied this ancient wisdom within the frontier of modern technology. They introduced a strategy that leverages multiple AI systems to discuss and argue with each other to converge on a best-possible answer to a given question. This method empowers these expansive language models to heighten their adherence to factual data and refine their decision-making. 

The crux of the problem with large language models (LLMs) lies in the inconsistency of their generated responses, leading to potential inaccuracies and flawed reasoning. This new approach lets each agent actively assess every other agent’s responses, and uses this collective feedback to refine its own answer. In technical terms, the process consists of multiple rounds of response generation and critique. Each language model generates an answer to the given question, and then incorporates the feedback from all other agents to update its own response. This iterative cycle culminates in a final output from a majority vote across the models' solutions. It somewhat mirrors the dynamics of a group discussion — where individuals contribute to reach a unified and well-reasoned conclusion.

One real strength of the approach lies in its seamless application to existing black-box models. As the methodology revolves around generating text, it can also be implemented across various LLMs without needing access to their internal workings. This simplicity, the team says, could help researchers and developers use the tool to improve the consistency and factual accuracy of language model outputs across the board.

“Employing a novel approach, we don’t simply rely on a single AI model for answers. Instead, our process enlists a multitude of AI models, each bringing unique insights to tackle a question. Although their initial responses may seem truncated or may contain errors, these models can sharpen and improve their own answers by scrutinizing the responses offered by their counterparts," says Yilun Du, an MIT PhD student in electrical engineering and computer science, affiliate of MIT CSAIL, and lead author on a new paper about the work . "As these AI models engage in discourse and deliberation, they're better equipped to recognize and rectify issues, enhance their problem-solving abilities, and better verify the precision of their responses. Essentially, we're cultivating an environment that compels them to delve deeper into the crux of a problem. This stands in contrast to a single, solitary AI model, which often parrots content found on the internet. Our method, however, actively stimulates the AI models to craft more accurate and comprehensive solutions."

The research looked at mathematical problem-solving, including grade school and middle/high school math problems, and saw a significant boost in performance through the multi-agent debate process. Additionally, the language models showed off enhanced abilities to generate accurate arithmetic evaluations, illustrating potential across different domains.

The method can also help address the issue of "hallucinations" that often plague language models. By designing an environment where agents critique each other's responses, they were more incentivized to avoid spitting out random information and prioritize factual accuracy. 

Beyond its application to language models, the approach could also be used for integrating diverse models with specialized capabilities. By establishing a decentralized system where multiple agents interact and debate, they could potentially use these comprehensive and efficient problem-solving abilities across various modalities like speech, video, or text. 

While the methodology yielded encouraging results, the researchers say that existing language models may face challenges with processing very long contexts, and the critique abilities may not be as refined as desired. Furthermore,the  multi-agent debate format, inspired by human group interaction, has yet to incorporate the more complex forms of discussion that contribute to intelligent collective decision-making — a crucial area for future exploration, the team says. Advancing the technique could involve a deeper understanding of the computational foundations behind human debates and discussions, and using those models to enhance or complement existing LLMs. 

"Not only does this approach offer a pathway to elevate the performance of existing language models, but it also presents an automatic means of self-improvement. By utilizing the debate process as supervised data, language models can enhance their factuality and reasoning autonomously, reducing reliance on human feedback and offering a scalable approach to self-improvement," says Du. "As researchers continue to refine and explore this approach, we can get closer to a future where language models not only mimic human-like language but also exhibit more systematic and reliable thinking, forging a new era of language understanding and application."

"It makes so much sense to use a deliberative process to improve the model's overall output, and it's a big step forward from chain-of-thought prompting," says Anca Dragan, associate professor at the University of California at Berkeley’s Department of Electrical Engineering and Computer Sciences, who was not involved in the work. "I'm excited about where this can go next. Can people better judge the answers coming out of LLMs when they see the deliberation, whether or not it converges? Can people arrive at better answers by themselves deliberating with an LLM? Can a similar idea be used to help a user probe a LLM's answer in order to arrive at a better one?"

Du wrote the paper with three CSAIL affiliates: Shuang Li SM '20, PhD '23; MIT professor of electrical engineering and computer science Antonio Torralba; and MIT professor of computational cognitive science and Center for Brains, Minds, and Machines member Joshua Tenenbaum. Google DeepMind researcher Igor Mordatch was also a co-author.

Share this news article on:

Related links.

• Joshua Tenenbaum
• Computer Science and Artificial Intelligence Laboratory (CSAIL)
• Department of Electrical Engineering and Computer Science

Related Topics

• Electrical engineering and computer science (EECS)
• Computer science and technology
• Artificial intelligence
• Computer vision
• Human-computer interaction

Related Articles

3 Questions: Jacob Andreas on large language models

3 Questions: How AI image generators could help robots

AI system makes models like DALL-E 2 more creative

Previous item Next item

More MIT News

Quantum repeaters use defects in diamond to interconnect quantum systems

Read full story →

Re-imagining the opera of the future

MIT welcomes nine MLK Visiting Professors and Scholars for 2023-24

Desalination system could produce freshwater that is cheaper than tap water

Improving US air quality, equitably

From physics to generative AI: An AI model for advanced pattern generation

• More news on MIT News homepage →

Massachusetts Institute of Technology 77 Massachusetts Avenue, Cambridge, MA, USA

• Map (opens in new window)
• Events (opens in new window)
• People (opens in new window)
• Careers (opens in new window)
• Accessibility
• Social Media Hub
• MIT on Facebook
• MIT on YouTube
• MIT on Instagram