AI Inventorship in Patent Law: Thaler v Comptroller of Patents (PDF)

bs_bs_banner Modern Law Review DOI: 10.1111/1468-2230.12907 ‘AI is not an Inventor’: Thaler v Comptroller of Patents, Designs and Trademarks and the Patentability of AI Inventions...

bs_bs_banner Modern Law Review DOI: 10.1111/1468-2230.12907 ‘AI is not an Inventor’: Thaler v Comptroller of Patents, Designs and Trademarks and the Patentability of AI Inventions Rita Matulionyte∗ The increasing use of Artificial Intelligence (AI) technologies in inventive processes raises nu- merous patent law issues, including whether AI can be an inventor under law and who owns the AI-generated inventions. The UK Supreme Court decision in Thaler v Comptroller of Patents, De- signs and Trademarks has provided an ultimate answer to this question: AI cannot be an inventor for the purposes of patent law. This note argues, first, that while such a human-centric approach to inventorship might discourage the use and development of AI technologies with autonomous invention capabilities, it will help retain an active human involvement in technologically sup- ported inventive processes and continuously foster human ingenuity. Second, despite the Court focusing on what patent law is and not on what the law should be, the decision will be influential in the ongoing discussions on the future of patent law and will make it more difficult to expand patent law to incorporate non-human inventors. Third, the decision has opened, or revealed, the gaps in patent law that the emergence of AI technologies have created and for which new legal solutions will be needed, especially with relation to the ownership of AI-assisted inventions and the validation of inventorship claims. INTRODUCTION The increasing use of sophisticated Artificial Intelligence (AI) technologies in conceptualising new inventions has led to an intense international debate on patent law implications in relation to innovations generated by or with the assistance of AI technologies.1 In particular, can AI generate inventions au- tonomously, ie without a significant human contribution? If so, are such AI- generated inventions protected under current patent laws, and should they be? ∗ Associate Professor at Macquarie Law School, Macquarie University, Australia. Open access publish- ing facilitated by Macquarie University, as part of the Wiley - Macquarie University agreement via the Council of Australian University Librarians. 1 See public consultations conducted by governments: UK Intellectual Property Office, ‘Govern- ment response to call for views on artificial intelligence and intellectual property’ (updated on 23 March 2021) at https://www.gov.uk/government/consultations/artificial-intelligence-and- intellectual-property-call-for-views/government-response-to-call-for-views-on-artificial- intelligence-and-intellectual-property [https://perma.cc/6LFP-D5PH]; US Patent and Trade Mark Office, ‘Public Views on Artificial Intelligence and Intellectual Property Policy’ (October 2020) at https://www.uspto.gov/sites/default/files/documents/USPTO_AI-Report_2020- 10-07.pdf [https://perma.cc/WL9W-6H5Q]; WIPO, Revised Issues Paper on Intellectual Property Policy and Artificial Intelligence (Document code: WIPO/IP/AI/2/GE/20/1 REV, 29 May 29 2020) at https://www.wipo.int/meetings/en/doc_details.jsp?doc_id=499504 [https://perma.cc/GW4E-TPPZ]. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. (2024) 00(0) MLR 1–14 This is an open access article under the terms of the Creative Commons Attribution License,which permits use,distribution and reproduction in any medium, provided the original work is properly cited.  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Patentability of AI Inventions Who could and should be recognised as an inventor and owner of such patents? Stakeholders, policy makers, courts and academics around the world have been actively examining these issues.2 In parallel, a similar discussion is happening in other areas of IP law, such as copyright, with no international consensus on whether creations generated with assistance from AI could and should be afforded copyright protection.3 The importance of these issues is unquestion- able: the decision to award or not award patent or copyright law protection for AI-generated outputs is likely to affect the future utilisation of AI technolo- gies in innovation and creative processes as well as the further expansion of AI technology sectors around the globe.4 The recent UK Supreme Court decision in Thaler v Comptroller of Patents, De- signs and Trademarks,5 which is a part of an international Artificial Inventor project led by Dr Stephen Thaler and his legal team,6 has provided an ultimate clari- fication on some of the questions regarding the patentability of AI-generated inventions, at least in the UK.7 Namely, the Supreme Court unanimously con- cluded that, under the Patents Act 1977, an AI system cannot be listed as an inventor in the patent application, and that the owner of the AI system that arguably generated the invention is not eligible to apply for a patent in this invention or to be granted a patent over it.8 This outcome is consistent with the outcomes of most other patent applications that Dr Thaler submitted in other jurisdictions and reinforces an emerging international consensus that AI cannot be an inventor for the purposes of current patent law.9 2 See for example Ryan Abbott, ‘I Think, Therefore I Invent: Creative Computers and The Future Of Patent Law’ (2016) 57 Boston College Law Review 1079; Ben Hattenbach and Joshua Glucoft, ‘Patents In An Era Of Infinite Monkeys And Artificial Intelligence’ (2015)19 Stanford Technology Law Review 32; for policy documents see ibid. 3 See for example Courtney White and Rita Matulionyte, ‘Artificial Intelligence Painting a Larger Picture on Copyright’ (2020) 30 Australian Intellectual Property Review 224; Russ Pearlman, ‘Rec- ognizing Artificial Intelligence (AI) as Authors and Inventors Under U.S. Intellectual Property Law’ (2018) 24 Richmond Journal of Law and Technology 1; Ana Ramalho, ‘Will Robots Rule the (Artistic) World? A Proposed Model for the Legal Status of Creations by Artificial Intelligence Systems’ (2017) 21 Journal of Internet Law 12; Tim W. Dornis, ‘Artificial Creativity: Emergent Works and the Void in Current Copyright Doctrine’ (2020) 22 Yale J L & Tech 1; Rita Mat- ulionyte and Jyh-An Lee, ‘Copyright in AI-generated works: lessons from recent developments in patent law’ (2022) 19 SCRIPTed – A Journal of Law, Technology and Society 5. 4 Similarly, Abbott, n 2 above, 1080. 5 Thaler v Comptroller of Patents, Designs and Trade Marks UKSC 49 (Thaler (SC)). 6 Information about the project can be found at https://artificialinventor.com/patent/ [https:// perma.cc/CPT4-S3N4]. 7 Note that after the Supreme Court judgment the applicant submitted a divisional application to the UK Patent Office No. 2206827.4, see ibid. 8 Thaler (SC) n 5 above per Kitchen LJ, with whom Hodge, Hamblin, Leggatt and Richards LLJ agreed. 9 See for example European Patent Office decision J 0008/20 (Designation of inventor/DABUS) 21-12-2021; Germany: Federal Patent Court decision re patent applications 11 W (pat) 5/21 and 18 W (pat) 28/20 (denies AI as inventor but allows listing Dr Thaler as an inventor and mentioning DABUS involvement); US: Thaler v Vidal, 43 F.4th 1207 (Fed. Cir. 2022); Australia: Commissioner of Patents v Thaler FCAFC 62; Israel: applications nos 268604 and 268605. The only jurisdiction that issued the patent for an invention generated by DABUS is South Africa, application number ZA2021/03242, for a comment see Desmond Osaretin Oriakhogba, ‘DABUS gains territory in South Africa and Australia: Revisiting the AI-inventorship question’ (2021) 9 South African Intellectual Property Law Journal 87. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. 2 (2024) 00(0) MLR 1–14  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Rita Matulionyte The next part of this note provides the factual background of Thaler v Comptroller of Patents, Designs and Trademarks and briefly reviews the decisions of the UK Intellectual Property Office (UKIPO), the High Court, the Court of Appeal and the UK Supreme Court. The following three parts focus on three main issues examined by the Supreme Court: i) whether AI can be listed as an inventor; ii) whether the AI owner can apply for and be granted a patent for an invention autonomously generated by a machine; and iii) whether the UKIPO can and should assess the truthfulness of the inventorship statement. The note concludes that the human-centric approach to inventorship, as adopted by the Supreme Court, is well-reasoned and maintains the internal consistency of the Patents Act 1977. While it might disincentivise the use and development of AI technologies with autonomous inventive possibilities, it helps ensure active human involvement in technologically-supported innovation and supports hu- man ingenuity more generally. Further, despite the decision focussing on current patent law only, it will make it more difficult to advocate for future patent laws that include non-human inventors. In addition, the decision helps reveal chal- lenges to and gaps in patent law that have emerged as a result of AI technologies, such as determining the ownership of AI-assisted inventions and ensuring the truthfulness of inventorship claims when AI is used in the inventive process. Such issues will require legal intervention and redress by law makers heading into the future as the use of AI is expected to proliferate. FACTUAL AND LEGAL BACKGROUND In 2018, Dr Stephen Thaler filed two patent applications in the UKIPO: one for an invention entitled ‘Food container’, claiming a fractal drink container that can change shapes, and the second for an invention claiming a flickering light that mimics brain activity – or ‘neural flame’.10 Subsequently, these applications were extended to other member states of the Patent Cooperation Treaty, in- cluding the US, Germany, Europe (ie European Patent Office), Australia, South Korea, Japan, Israel, Canada and New Zealand.11 The statement of inventorship lodged with the UKIPO on 24 July 2019 nominated an AI system ‘DABUS’ as the inventor.12 Dr Thaler argued that he had a right to apply for and be granted a patent ‘by ownership of the creativity machine “DABUS’”.13 Decisions of UKIPO, High Court and Court of Appeal The UKIPO concluded, first, that ‘DABUS is not a person as envisaged by sections 7 and 13 of the [Patent] Act and so cannot be considered an inventor’. 10 Patent application nos GB1816909.4 and GB1818161.0. 11 The up-to-date list of applications and their outcomes is available at ‘Artificial Inventor’ website, n 6 above. 12 DABUS stands for ‘Device for the Autonomous Bootstrapping of Unified Sentience’. 13 Thaler v Comptroller General of Patents, Designs and Trade Marks EWCA Civ 1374 (Thaler (CA)) at. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. (2024) 00(0) MLR 1–14 3  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Patentability of AI Inventions Secondly, the UKIPO found that ‘the applicant is … not entitled to apply for a patent simply by virtue of ownership of DABUS, because a satisfactory derivation of right has not been provided’.14 For these reasons, the UKIPO concluded that the applications were deemed to be withdrawn under section 13 of the Patents Act 1977.15 Dr Thaler’s appeal to the High Court was dismissed by Marcus Smith J who agreed with all three reasons provided by the UKIPO.16 Similarly, the majority in the Court of Appeal,17 Arnold and Laing LJJ, rejected Dr Thaler’s appeal on three grounds: first, DABUS could not be an inventor because it was not a person; secondly, Dr Thaler had not established a right to be granted the patent as set in section 7(2) of the Patents Act 1977; and, thirdly, the Deputy Director of the UKIPO had been correct to find that the applications were deemed withdrawn because the requirements of section 13(2) of the Patents Act 1977 to identify the inventor and the derivation of a right to apply for a patent were not satisfied. In a dissenting opinion, Birss LJ concurred that machines do not qualify as ‘inventors’ under the Patents Act 197718 but disagreed with the majority’s interpretation of section 13(2). Birss LJ concluded that ‘[i]t is no part of the Comptroller’s functions under the 1977 Act to in any way be satisfied that the applicant’s claim to the right to be granted the patent is good. In granting a patent to an applicant the Comptroller is not ratifying the applicant’s claim to derivation’.19 According to Birss LJ, the fact that the machine was named as an inventor of the claimed invention was not an adequate ground to consider the application withdrawn under section 13(2). UK Supreme Court decision The UK Supreme Court, comprising Lord Hodge, Lord Kitchin, Lord Ham- blen, Lord Leggatt and Lord Richards, handed down their 99-paragraph deci- sion on 20 December 2023, which unanimously rejected an appeal from Dr Thaler.20 The Supreme Court examined the scope and meaning of the concept of ‘inventor’ in the Patents Act 1977 and agreed with the earlier findings that only a natural person can be an inventor, and not a machine. They based this con- clusion on a contextual interpretation of the Patents Act 1977, and in particular sections 7 and 13. Section 7(3) of the Patents Act 1977 defines ‘inventor’ as ‘the actual deviser of the invention’. While the definition does not explicitly refer to a ‘person’ or a ‘human’, the Supreme Court concluded that ‘[t]here is no 14 UK IPO decision BL O/741/19 at. 15 ibid. 16 Thaler v Comptroller General of Patents, Designs and Trade Marks EWHC 2412 (Pat); for a summary of the High Court decision see Enrico Bonadio, Luke McDonagh and Plamen Dinev, ‘Artificial intelligence as inventor: exploring the consequences for patent law’ (2021) 1 IPQ 48, 59-60. 17 Thaler (CA) n 13 above. 18 ibid at -. 19 ibid at (iv). 20 Thaler (SC) n 5 above. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. 4 (2024) 00(0) MLR 1–14  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Rita Matulionyte suggestion that ‘deviser’ here has anything other than its ordinary meaning, that is to say, a person who devises a new and non-obvious product or process (the invention) which is capable of industrial application and may be protected un- der the patent system’.21 The Supreme Court referred to the House of Lords’ decision in Rhone-Poulenc Rorer International Holdings Inc v Yeda Research and Development Co Ltd where Lord Hoffmann explained that ‘the inventor’, under section 7(3) of Patents Act 1977, is a natural person who came up with the in- ventive concept,22 the statement previously approved by Laddie J in University of Southampton’s Applications.23 The Supreme Court also examined a broader con- text of the Patents Act 1977, and in particular sections 2(4), 8, and 37 of the Act, which explicitly contemplate that the inventor is a natural person, that rights to apply for and be granted a patent are derived through the inventor only, and that patents are only granted or transferred to persons with legal personality.24 After deciding that only a human can be an inventor under the Patents Act 1977, the Supreme Court examined whether Dr Thaler was nevertheless enti- tled to apply for and obtain a patent in respect of any technical advance made by DABUS.25 As in previous instances, the applicant relied on the accession doctrine, which generally suggests that the owner of the tangible property (for example an apple tree) also owns the tangible produce of this property (for ex- ample apples).26 By analogy with an application of this doctrine, the applicant contended that his ownership of the DABUS system resulted in him owning the invention generated by DABUS and thus the right to apply for and be granted patents for these inventions.27 The DABUS inventions are the ‘fruits’ of the DABUS machine, in the sense that they were produced by the machine. Accordingly, the applicant arguably satisfied the terms of section 7(2)(b) of the Act as he derived the title to apply for a patent ‘by operation of a rule of law’ – namely, the doctrine of accession.28 In response to that argument, the Supreme Court reiterated that section 7(2) is premised on the idea that ‘there must be an inventor, and that inventor must be a person.’29 Thus the references to ‘the invention’ in section 7(2)(b) on which Thaler relied are necessarily references to an invention devised by a person.30 The machine cannot devise any invention, but rather it would be a mere ‘tech- nical advance’.31 The Supreme Court further found that while the accession doctrine has been traditionally applied to tangible property, it has not been applied to intangible items, and there is no principled basis for applying the doctrine of accession in the given circumstances.32 21 ibid at (emphasis in original). 22 Rhone-Poulenc Rorer International Holdings Inc v Yeda Research and Development Co Ltd UKHL 43; Bus LR 1796 at. 23 University of Southampton’s Applications EWHC 2107 (Pat); RPC 220, 234. 24 Thaler (SC) n 5 above at -. 25 ibid at. 26 ibid at ; Blackstone’s Commentaries on the Laws of England (Clarendon Press, 1766) Book II, Chapter 26, 404-405 at. 27 Thaler (SC) ibid at. 28 ibid at. 29 ibid at. 30 ibid at. 31 ibid at. 32 ibid at. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. (2024) 00(0) MLR 1–14 5  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Patentability of AI Inventions The final issue that the Supreme Court addressed was whether the UKIPO was entitled to hold that the applications would be taken to be withdrawn for failure to comply with section 13 of the Patents Act 1977. Section 13(2) of the Act requires the applicant for a patent to indicate the person(s) whom the applicant believes to be the inventor(s) and, where the applicant is not an inventor, how the right to be granted the patent was derived. The applicant submitted that he had indeed listed an inventor (as required by the section), and the Hearing Officer was not permitted to investigate whether the claims were true, a proposition with which Birss LJ agreed in his dissenting judgment in the Court of Appeal.33 The Supreme Court agreed that it is no part of the function of the Comptroller to examine the correctness of genuine and plausible statements of inventorship and entitlement under section 13(2).34 Nonetheless, section 13 contains several requirements, one of them being to identify a person or persons who are believed to be the inventor/s of the claimed invention. Dr Thaler’s as- sumption that an inventor can be a non-human was shown to be incorrect, and he did not make any mistake of fact in identifying the correct inventor. Thus, by choosing to list the AI system as an inventor, which is legally impermissible, Dr Thaler failed to satisfy section 13(2)(a).35 In addition, Dr Thaler also failed to provide sufficient information about the derivation of rights required under section 13(2)(b). The Supreme Court held that whilst it is not the function of the UKIPO to investigate the correctness of an apparently genuine statement regarding the derivation of the applicant’s right to be granted the patent,‘that does not mean to say that the UKIPO is powerless to intervene where that indication is obviously defective or insufficient.’36 As a result, Dr Thaler’s assertion that he owned DABUS was found to be insufficient to satisfy section 13(2)(b) of the Patents Act 1977,and,accordingly,the UKIPO’s decision to consider the application withdrawn had been correct. HUMAN-CENTRIC APPROACH TO INVENTORSHIP UNDER CURRENT AND FUTURE PATENT LAW By concluding that the Patents Act 1977 in its current shape is premised on human inventorship and thus allows only humans to be recognised as inventors, the Supreme Court delivered a logical, well-reasoned decision and maintained the internal consistency of the Act. This decision, if left unchanged by law makers, is likely to impact how companies use AI in the inventive process and, in turn, the development of AI technologies with inventive capabilities. Further, despite focusing on current patent law only, the decision is likely to play a significant role in the discussions around the future of patent law, as will be explained below. 33 See discussion above. 34 Thaler (SC) n 5 above at. 35 ibid at -. 36 ibid at. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. 6 (2024) 00(0) MLR 1–14  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Rita Matulionyte Human-centric inventorship concept under the Patents Act 1977 The Supreme Court’s conclusion that only humans can be inventors was based on the contextual interpretation of the ‘inventor’ concept under the Patents Act 1977, and on the Court’s attempt to keep the internal consistency of the Act.37 After first analysing the meaning of ‘inventor’ within the scheme of section 7, the Court then examined it within the broader context of the Act, specif- ically referring to provisions that explicitly refer to a ‘person’.38 Indeed, this analysis helped to reach a conclusion that maintains the internal consistency of the Act. If an ‘inventor’ was defined in a broad way, as any devisor of the invention and not necessarily as a human devisor, some sections of the Patents Act 1977 that do not specifically refer to a person would apply to any inven- tor, both human and non-human, while others that explicitly refer to a person (such as sections 2(4), 8 and 37), would apply to human inventors only. While differentiated treatment of non-human and human inventions can be justified in certain circumstances,39 unequal treatment of human and non-human in- ventors resulting from a broad interpretation of the ‘inventor’ concept in the Patents Act 1977 would be inadvertently anomalous rather than the product of a well thought-through decision on the part of the policy makers. The lack of a clear rationale for such unequal treatment would have required legislative intervention to correct the inconsistences arising from a broad interpretation of the ‘inventor’ concept.40 If such a human-centred approach to the inventorship concept is retained in patent law, it is likely to influence how companies use AI in inventive processes. Since the use of AI in the inventive process might affect the patentability of the inventions, this might reduce incentives to use, and consequently develop, AI tools that are capable of autonomous inventions.41 On the other hand, it is not clear whether such autonomous inventions are desirable in the first place. Arguably, they might lead to a flood of AI-made inventions and patent applica- tions, which would stifle innovation.42 Keeping in mind that the most advanced inventive AI tools are likely to be in the hands of large corporations, allowing the patenting of AI-generated inventions might lead to the concentration of inventions in the hands of a small number of companies and further increase 37 See Thaler (SC) n 5 above, references to ‘consistent’ at , , ,. 38 Patents Act 1977, ss 2(4), 8, 37. 39 Numerous authors suggest that AI-generated inventions should be treated differently from human-generated inventions for various reasons, see for example Robert Plotkin, The Genie in the Machine: How Computer-Automated Inventing is Revolutionizing Law & Business (Stanford, CA: Stanford University Press, 2009); Shlomit Yanisky-Ravid and Xiaoqiong (Jackie) Liu, ‘When ar- tificial intelligence systems produce inventions: An alternative model for patent law at the 3a era’ (2018) 39 Cardozo Law Review 2215 (advocating non-patent incentives for AI innovation). 40 For example under Patents Act 1977, s 7(2)(a), a patent will be granted primarily to the inventor or joint inventors; however, since AI does not have a legal personhood, it could not become an owner of a patent. 41 Similarly, Ryan Abbott, ‘Hal the Inventor: Big Data and Its Use by Artificial Intelligence’ in Cassidy R. Sugimoto, Hamid R. Ekbia and Michael Mattioli (eds), Big Data Is Not a Monolith (Cambridge, MA: MIT Press, 2016) 1 at https://papers.ssrn.com/sol3/papers.cfm?abstract_id= 2565950 [https://perma.cc/X5EU-XWCK]; Abbott, n 2 above. 42 Erica Fraser, ‘Computers as inventors - legal and policy implications of artificial intelligence on patent law’ (2016) 13 SCRIPTed 305, 322. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. (2024) 00(0) MLR 1–14 7  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Patentability of AI Inventions the problem of patent thickets.43 Also, the requirement of human inventorship under patent law will help ensure that humans remain in charge of the inventive process. AI threatens to deskill certain qualified workers, so requiring humans to maintain active involvement in the inventive process in order to gain inventor- ship would allow for the preservation and continuing development of essential skills in specific industries and further encourage human ingenuity.44 Impact on the future of patent law While the decision concerns what patent law is and not what the law should be, it will nevertheless affect the ongoing policy discussions on the future of patent law.45 The Supreme Court expressly stated that they are examining what the inventor concept currently means, but not examining whether AI should be recognised as an inventor under the Patents Act 1977 in the future.46 In addi- tion, the Court avoided the application of a purposive approach,47 a statutory interpretation approach broadly used by courts.48 The Court briefly mentioned the parliamentary intention of the Act by merely supporting the conclusion of Laing LJ in the Court of Appeal that ‘Parliament did not have [machines] in mind when enacting this scheme’,49 but did not engage with the goals of the Patents Act 1977 in more depth. While reasons for this might be different, such as the lack of expressly stated goals in the Patents Act 1977,50 it is likely that the Court did not engage with the goals and purposes of the Act due to its politically charged nature. Indeed, the purposes or goals of the Patents Act 1977, and patent law more generally, have been a focus of recent academic and policy discussions: is patent law meant to encourage human innovation only or any innovation that has po- tential benefits to society? Some commentators have argued that the ultimate aim of patent law is to encourage innovation that could benefit the public, de- spite who or what generates a new idea.51 While machines do not require any 43 Michael J. Meurer, ‘Business Method Patents and Patent Floods’ (2002) 8 Washington University Journal of Law & Policy 309. 44 See Fraser, n 42 above, 323-324; Plotkin, n 39 above, 107, 136; see also Ravid and Liu, n 39 above. 45 Note that the Brazilian Parliament member proposed a Bill which allows listing AI as both an inventor and owner of the patent: Roberto Rodrigues Pinho and others, ‘Brazilian Con- gressman introduces bill to allow AI as inventor’ Kluwer Patent Blog 26 February 2024 at https://patentblog.kluweriplaw.com/2024/02/26/brazilian-congressman-introduces-bill-to- allow-ai-as-inventor/#:∼:text=Bill%20%23303%2F2024%20proposes%20the,the%20 artificial%20intelligence%20system%20that [https://perma.cc/7BTP-A4HJ]. 46 Thaler (SC) n 5 above at -. 47 For an overview of the purposive construction see Oliver Jones, Bennion on Statutory Interpretation (London: Lexis Nexis, 6th ed, 2017) 845-868. 48 See for example A-G’s Reference (No 5 of 2002) UKHL 40; 4 All ER 901 at ; Kammins Ballrooms Co Ltd v Zenith Investments (Torquay) Ltd AC 850, 880. 49 Thaler (SC) n 5 above at. 50 The introduction to the Patent Act 1977 refers to the intentions of the Act but not to its public policy goals (‘An Act to establish a new law of patents applicable to future patents and applications for patents; to amend the law of patents applicable to existing patents and applications for patents; to give effect to certain international conventions on patents; and for connected purposes’). 51 Abbott, n 41 above, 1; Abbott, n 2 above; Fraser, n 42 above, 325-326; Hattenbach and Glucoft, n 2 above, 32. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. 8 (2024) 00(0) MLR 1–14  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Rita Matulionyte incentives to generate inventions,52 humans need incentives to develop such in- ventive machines,and patents over machine-generated inventions could provide such an incentive.53 In contrast, others have argued that the aim of patent law has always been to promote human innovation,54 and there is no sufficient ratio- nale or economic reason to award patents for computer-generated inventions.55 According to some commentators, AI merely performs the algorithm designed by humans and is thus not capable of independent invention.56 If AI-generated inventions could be patented, this would arguably lead to numerous unintended consequences, such as a possible flood of patent applications, a concentration of inventions in the hands of large companies and a negative impact on human ingenuity more generally.57 It is likely that the Court kept in mind the absence of academic and political consensus on the purposes of patent law and avoided commenting on them in order not to interfere with these politically charged discussions. Despite that, the Supreme Court’s decision, and especially the establishment of human inventorship as a core premise of patent law, will influence ongoing policy discussion on the future of patent law. The Court’s conclusion that the Patents Act 1977 was designed and adopted without machines in mind,58 and has been traditionally premised on human inventorship,59 will now form a basis for this discussion.Those in favour of recognising AI as a possible inventor under patent law now have a burden to convince law makers why some of the under- lying premises of patent law, such as human inventorship, need to be changed. They will have to prove that new technological circumstances require an over- haul to the patent system as it has been known, and that including non-human inventors will benefit not only companies using AI but the entire society. If the Court had agreed with the broad interpretation of ‘inventor’ which includes non-human inventors, the burden of proof would have been on the opponents of this approach to convince law makers that patent law needs to be changed to exclude non-humans from the inventorship concept. Thus, while the deci- sion to reject AI as a possible inventor and confirm human inventorship as a fundamental premise of patent law is not revolutionary, the significance of the Supreme Court findings should not be underestimated, as they will affect both the application of the current Patents Act 1977 and the development of future patent law. 52 Mark Perry and Thomas Margoni, ‘From Music Tracks to Google Maps: Who Owns Computer- generated Works?’ (2010) 26 Computer Law and Security Review 621, 627. 53 Fraser, n 42 above, 326; Pearlman, n 3 above, 23; see also W. Michael Schuster, ‘Artificial intel- ligence and patent ownership’ (2018) 75 Washington and Lee Law Review 1945 (on allocation of ownership of AI-generated inventions). 54 Eva Stankova, ‘Human inventorship in European patent law’ (2021) 80 Cambridge Law Journal 338. 55 For example Hattenbach and Glucoft, n 2 above, 50. 56 Daria Kim, ‘AI-Generated Inventions: Time to Get the Record Straight?’ (2020) 69 GRUR International 443; see also Joseph Straus, ‘Artificial intelligence and patenting: some lessons from “DABUS” patent applications’ (2022) 44 E.I.P.R. 348, 354 (‘The facts offered by Thaler do not seem to support his claim that DABUS autonomously invented the claimed invention’). 57 Fraser, n 42 above, 323-324; Meurer, n 43 above; Plotkin, n 39 above, 107, 136; see also Ravid and Liu, n 39 above. 58 Thaler (SC) n 5 above at. 59 ibid (human inventorship as ‘fundamental premise of 1977 Act’). © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. (2024) 00(0) MLR 1–14 9  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Patentability of AI Inventions PATENT RIGHTS CAN ONLY BE DERIVED FROM A HUMAN INVENTOR The Supreme Court’s findings that the accession doctrine does not apply to intangibles and thus a right to apply for a patent can be derived from humans only should be welcomed. This is consistent with a human-centric approach to inventorship, as adopted by the Court, and avoids the unreasonable expansion of patent and intellectual property law more generally. At the same time, this conclusion has created, or exposed – but did not resolve – the gaps that have emerged in patent law as a result of the emergence of AI technologies. The decision that, under sections 7(2)(b) and (c) of the Patents Act 1977, persons can derive the right to be granted a patent only from another person, and not from any non-human entity such as a machine, was achieved by de- clining to extend the accession doctrine to intangibles, such as a right to apply for and be granted a patent.60 Indeed, if the Court had decided to extend the application of the accession doctrine to intangibles, this would have significantly – and without a sound justification – both broadened the scope of the doctrine itself and the scope of protection granted by intellectual property rights more generally. If an owner of a tangible thing, such a machine, acquired exclusive rights (property or similar) over intangible outputs produced by the machine, such as new concepts, this would go far beyond the ability of such owners to apply for a patent over such concepts. The accession doctrine, if extended, would then apply to other intan- gibles too, such as copyright. As a result, if a machine produces an image, text, or a musical work without significant human intervention (ie autonomously), the owner of the machine would be the owner of copyright to the work. This is against current requirements for human authorship under copyright law in most jurisdictions.61 A similar conclusion would also apply to the outputs of other non-humans that humans own, such as pets: the owner of a pet would own the copyright over paintings, music, or performances by pets. Works cre- ated by animals have been traditionally denied copyright protection,62 and thus the extension of the accession doctrine to intangibles would lead to a signifi- cant expansion of copyright law. Such an expansion of intellectual property law would lack a sound justification.63 60 For an explanation of the accession doctrine, see the subsection headed ‘UK Supreme Court decision’ above. 61 See US:Thaler v Perlmutter,1:22-cv-01564;Australia:ceTV Pty Ltd v Nine Network Australia Pty Ltd HCA 14; exceptionally, the UK has sui generis protection for computer-generated works with rights vesting to a person who made ‘necessary arrangements’ (Copyright, Designs and Patents Act 1988, s 8). However, it remains unclear how these provisions apply to AI-generated works – see Jyh-An Lee, ‘Computer-Generated Works under the CDPA 1988’ in Jyh-An Lee, Reto Hilty and Kung-Chung Liu (eds), Artificial Intelligence and Intellectual Property (Oxford: OUP, 2021). 62 Naruto v Slater et al 888 F.3d 418 (9th Cir. 2018) (confirming that the US Copyright Act did not expressly authorise animals to file copyright infringement suits). 63 For example Patrick Zurth, ‘Artificial Creativity? A Case against Copyright Protection for AI- Generated Works’ (2020-2021) 25 UCLA J L & Tech i; Jane C. Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Tech L J 343. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. 10 (2024) 00(0) MLR 1–14  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Rita Matulionyte On the other hand, the conclusion that patent rights can initially be derived from humans only will have unintended consequences, or create new gaps in the law, especially with relation to inventions generated with the assistance of AI (AI-assisted inventions). If a human is using AI in an inventive process, and both a human and an AI system contribute inventive concepts to the final invention, how would the inventorship and ownership of the entire invention be determined? In a case where both inventors were human, they would qualify as joint inventors. When one inventor is AI, AI cannot be a joint inventor because, according to the Supreme Court, only a human can be an inventor. In this situation, how will the law treat the inventive contribution made by an AI system? One option is to distinguish it from a human contribution and exclude or disclaim it from a patent application and a subsequent patent. A similar option has been chosen by the US Copyright Office, which requires authors to disclose the use of AI and disclaim from copyright protection the parts of work created by AI.64 This approach could be criticised for its impracticality, especially when iterative use of AI throughout the creative process makes it impossible or un- feasible to meaningfully distinguish the human and non-human contributions, and the subsequent difficulty of exercising the rights over a certain part of the work only.65 An alternative solution is to allow the human inventor, or another entitled person, to claim a patent for the entire invention, ie including the part of the invention generated by AI. A legal basis would be needed for such a solution and, if the accession theory applied to intangibles, it could have served as such a potential legal basis. According to such extended accession theory, the human inventor or other entitled person (for example employer) who owns AI used in the inventive process would be able to claim a right to the part of the invention generated by AI, even if they did not invent that part (or the entirety) of the invention. After the Supreme Court rejected the accession doctrine as a ground to acquire a right into a technical advancement generated by AI, it closed the door for this legal solution. Courts or law makers will need to search for an- other legal basis which will enable the applicant to establish ownership over AI-assisted inventions in their entirety, ie including the part invented by AI.66 64 US Copyright Office, ‘Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence’ 16 March 2023, 88 FR 16190, 16192. 65 See for example US Copyright Review Board, ‘Second Request for Reconsidera- tion for Refusal to Register Théâtre D’opéra Spatial (SR # 1-11743923581; Corre- spondence ID: 1-5T5320R)’ 5 September 2023 at https://fingfx.thomsonreuters.com/gfx/ legaldocs/byprrqkqxpe/AI%20COPYRIGHT%20REGISTRATION%20decision.pdf [https: //perma.cc/HJT5-H86F]. While rejecting protection for AI-generated elements of image, the Board opened a possibility that edits done by a human could be protected by copyright; it did not discuss how rights into such edits could be exercised. 66 For example in early 2024, the US Patent and Trademark Office issued ‘Inventorship Guid- ance for AI-assisted inventions’ at https://public-inspection.federalregister.gov/2024-02623.pdf? 1707745519 [https://perma.cc/Z8BA-S2YJ]. The Guidance, however, does not deal with this question. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. (2024) 00(0) MLR 1–14 11  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Patentability of AI Inventions UKIPO DUTY (NOT) TO ASSESS THE ENTITLEMENT TO APPLY FOR A PATENT Finally, the Supreme Court decision in relation to section 13, though justified, raises further questions about the adequacy of section 13 and other provisions under the Patents Act 1977 in ensuring the truthfulness of inventorship claims, especially in cases when AI is used in the inventive process. The conclusion that Dr Thaler’s disclosure of the inventor and the deriva- tion of rights was deficient and did not meet the requirements of section 13 of the Patents Act 1977 was consistent with the main conclusion of the deci- sion, namely, that the Act is premised on human inventorship. However, it is questionable whether the extent of rights given to the Patent Office to assess compliance with section 13 is sufficient in light of the increasing use of AI in inventive processes.67 The question that becomes relevant with the use of AI technologies by in- dividual inventors or companies is who could question whether the human or humans indicated as inventors in the patent application are the actual inventors (ie the nominated inventor(s) and not a machine) that devised or sufficiently contributed to the invention claimed in the patent application, and on what grounds this could occur.68 With AI being able to significantly contribute to inventive concepts, and the Supreme Court rejecting the possibility of listing AI as an inventor, compa- nies using AI in the inventive process have little choice but to list humans as the only inventors of the inventions claimed, even if they only partially devised the invention, or minimally, insignificantly or only indirectly participated in the inventive process (for example by providing instructions to the AI system). Currently, the inventorship and ownership of a patent can be challenged both before and after the patent is granted but, usually, only by a person who claims to be the actual inventor or owner of that patent.69 In a case where the inven- tion was entirely or partially generated by AI and the applicant who used AI in the inventive process lists themselves or other natural persons (for example employees) as inventors, no other person will be eligible to challenge the in- ventorship or ownership of the patent since no other person could legitimately claim that they are the inventor or owner of the invention or of the subsequent patent. The Comptroller’s right to revoke the patent is very limited too, and incorrect inventorship or ownership is not among the grounds on which the Comptroller could initiate the patent revocation procedure.70 While the Supreme Court provided the Comptroller with certain rights to review the inventorship and ownership claims, this ability applies only when the claims are ‘defective or insufficient’, for example when a non-human is 67 For an overview of AI use in inventive processes see Fraser, n 42 above, 314-319. 68 This is to some extent considered in the US Patent and Trademark Office,‘Inventorship Guidance for AI-assisted inventions’ n 66 above. 69 See for example Patents Act 1977, s 8 (any person who considers they are entitled); s 37 (‘any person having or claiming a proprietary interest in or under the patent’); s 72(2)(a) (a person ‘entitled to be granted that patent’), compare with s 13(3) (refers to ‘any person’). 70 For example Patents Act 1977, s 73 (allows Comptroller to revoke a patent only in cases when ‘an invention for which a patent has been granted formed part of the state of the art’). © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. 12 (2024) 00(0) MLR 1–14  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Rita Matulionyte listed as an inventor.71 This power is unlikely to be sufficient in the context of AI-generated and AI-assisted inventions. Namely, if the invention was gener- ated by or with the assistance of AI but only humans were listed as inventors, the Comptroller would have no opportunity to challenge the statement on an inventorship and entitlement issue. First, they are unlikely to know that AI has conceived the invention or a part of it,because there is currently no requirement to disclose the use of AI in the inventive process.72 Secondly, if the applicant genuinely ‘believes’ that the persons listed as inventors sufficiently contributed to the invention and are thus eligible to be named as inventors, the statement is not ‘deficient or insufficient’ and thus the Comptroller does not have a power to further investigate the statement or conclude that the statement did not meet the requirements under section 13.73 Despite the Supreme Court doing what they could to make section 13 more flexible, the limited role of the Comptroller in clarifying the inventorship and entitlement to apply for a patent does not provide an adequate answer to the emerging – but real – problems caused by AI technologies. Under current rules, neither interested parties (for example competitors) nor the Comptroller can challenge the inventorship or ownership and thus the patentability of inventions generated with significant help from powerful AI systems. It will thus be the task of law makers to repair this new gap in the law that emerged as a result of the increasing use of AI technologies in inventive processes. If law makers de- cide to maintain a human-centered approach to patent law and only allow the patenting of inventions with sufficient human contributions, they will have to introduce avenues to validate whether the inventors claimed in the patent appli- cations are actual inventors and whether the inventions claimed were invented with sufficient human contribution. This might require the establishment of duties to disclose the use of AI in the inventive process,74 expanding a range of persons who are entitled to challenge the inventorship statements, and/or providing broader powers for the Comptroller to investigate the truthfulness of inventorship claims in certain situations.75 CONCLUSION The UK Supreme Court decision in Thaler v Comptroller of Patents, Designs and Trademarks is significant for at least three reasons. First, it concluded that the current Patents Act is premised on human inventorship, and thus inventions 71 Thaler (SC) n 5 above at. 72 Compare: the US Copyright Office has introduced such a disclosure requirement for works generated with assistance from AI, see n 64 above, 16192. 73 See Patents Act 1977, s 13(2)(a) (‘identifying the person or persons whom he believes to be the inventor or inventors’ – highlight by author). 74 See Bonadio, McDonagh and Dinev, n 16 above, 61. In contrast, no such specific disclosure duty was introduced in the recently issued US Patent and Trademark Office, ‘Inventorship Guidance for AI-assisted inventions’ n 66 above. 75 Notably, the UKIPO Guidelines to the examination of patent applications relating to AI inven- tions (published on 22 September 2022) have been suspended and are currently under reconsid- eration by the Office. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. (2024) 00(0) MLR 1–14 13  14682230, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/1468-2230.12907 by Luiss Lib University Degli Stu, Wiley Online Library on [03/10/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Patentability of AI Inventions generated by AI and without sufficient human contribution cannot be patented. While this might disincentivise the development of AI technologies with fully autonomous inventive capacities, it will help to maintain human involvement in innovation and further encourage human ingenuity. Second, even though the Supreme Court avoided a discussion on the goals of patent law and whether it should maintain a human-centric approach to inventorship in the future, this decision is likely to be influential in the ongoing policy discussions on the future of patent law. The decision will make it more difficult to introduce AI systems as possible inventors under patent law, with the advocates of this approach having to prove that a human-centric patent system no longer meets the needs of society. Third, the decision has created, or made obvious, new gaps in patent law resulting from the emerging use of AI in inventive processes. In the case of AI-assisted inventions, who can claim ownership of the part of inventions gen- erated by AI, and how can this claim be made? While the accession doctrine could have possibly assisted on this matter, the Supreme Court has rejected its application in relation to intangibles. Further, who will be able to question whether the inventions claimed in patent applications are indeed invented by the persons listed as inventors and not by AI tools, and how would this occur? Powers granted to the Comptroller under section 13 of the Patents Act 1977, as construed by the Supreme Court, will be insufficient for this purpose. These and other emerging gaps in patent law will need urgent legal solutions to en- sure that patent law does not discourage legitimate uses of AI technologies in inventive processes and impede technologically supported innovation. © 2024 The Author(s). The Modern Law Review published by John Wiley & Sons Ltd on behalf of Modern Law Review Limited. 14 (2024) 00(0) MLR 1–14