Unlocking the Canvas: A Legal Analysis of AI Patents in the US and UK

In the rapidly evolving realm of technology, generative Artificial Intelligence (AI) stands out as a revolutionary force, fundamentally altering industries and redefining our understanding of creativity. However, As AI experiences rapid growth and exerts a profound influence across various sectors, legal questions surrounding the patentability of AI-generated works have come to the forefront. This explosion in AI’s creative capabilities comes with a legal conundrum: can AI-generated art and ideas be subject to patent protection?

The cases of Thaler v. Perlmutter in the US and Thaler v. Comptroller-General of Patents, Designs and TradeMarks in the UK serve as pivotal cases that can be used to unravel the complex web of legislation and court precedents shaping the patentability of AI-generated creations. Upon examining these cases, it becomes evident that there is no compelling justification within both US and UK law for the exclusion of AI art and ideas from patent protection. This underscores the importance of governmental agencies reevaluating the use of AI tools.

The case of Thaler v. Perlmutter revolves around the question of whether AI-generated art should be eligible for copyright protection. The plaintiff, Stephen Thaler, sought copyright registration for a piece of art generated by his computer system, “Creativity Machine.”As the owner of the computer system, Mr. Thaler attempted to get copyright protection for the art by listing the system as the author and explaining that because he owns the system, the copyright should transfer to him. [2] The United States Copyright Office, under the direction of Shira Perlmutter, denied the registration, asserting that copyright protection was only available for works created by human authors. [3]

The court cited precedent all the way from the 1800s, focusing primarily on the language of the Copyright Acts of 1909 and 1976 for the arguments made. The court highlighted specifically the case of Burrow-Giles Lithographic Co. v. Sarony, in which it was determined that the amendment of the Copyright Act to cover photographs on books was constitutional. [4] The court, in citing this precedent, hoped to highlight that copyright malleability, [defined here], is an essential part of the Copyright Act. The court under Justice Howell discusses malleability at length, saying that, “Copyright is designed to adapt with the times.” Showing that there has been a consistent understanding that human creativity is the at the core of copyrightability, even as that human creativity is channeled through new tools or into new media.” [5] The court openly agrees that copyright evolves with new instruments of human creativity, but the current legal policies simply aren’t able to account for AI generated works.

This case challenges the conventional notion that copyright protection is reserved exclusively for human-authored works. In the case, Thaler argued  that because he owns the technology that created the work of art, he has authorship over it. The court, citing the Copyright Act of 1976, claims that copyright protection is provided to “original works of authorship… either directly or with the aid of a machine or device.” (102(a).) Further, the court says “the ‘fixing’ of the work in the tangible medium must be done ‘by or under the authority of the author.’” (Id. 101.). The court highlights that for a work to be eligible for copyright it must have an author, which the court later defines as being “[t]he creator of an artistic work; a painter, photographer, filmmaker, etc.” [6]

The policies and cases cited in Thaler v. Perlmutter stem from a time when the creative capabilities of generative AI technology were not as advanced as they are today. In its ruling, the court relies on precedents that were established in an era when the idea of AI generating original and creative works was more speculative than practical. Though the court admits the adaptability of copyright law, the wording of the Copyright Act (1976) and of precedent surrounding such copyrights remains quite ambiguous. [7]

[if it would be right] For example, the court cited Kelley v. Chicago Park District (7th cir. 2011), a case that emphasized the requirement of human authorship in claiming copyright. The court refused to recognize the copyright of a garden that Chapman Kelley  cultivated, claiming that the garden’s form was due to the forces of nature. This comes in spite of the clearly human-conceived plan for the arrangement of the garden. The court’s definition of human involvement further demonstrates itself to be ill-defined, as the court also cites Penguin Books U.S.A., Inc. v. New Christian Church of Full Endeavor, a case where a woman heard “voices” and wrote the words down into a book. Through later collaboration, she published the book with the words from the “voices”. [8] The court asserted that since the 'voices,' although serving as the source, were heard and recorded by a human, the work was deemed to have human origins. The court's reliance on precedents rooted in an entirely different landscape is not applicable to the complexities posed by AI-generated art, and hence require reevaluation of what authorship means.

The court's adherence to an ambiguous interpretation of human authorship fails to acknowledge the collaborative nature inherent in AI creation. In the realm of AI-generated art, humans significantly contribute by designing, training, and overseeing the AI systems. As articulated by H. James Wilson and Paul R. Daugherty in an article in the Harvard Business Review, the successful functioning of AI relies on sustained human involvement, providing it with the necessary information to execute its intended tasks. [9] Ignoring this collaborative aspect oversimplifies the intricate relationship between humans and AI within the creative process.

The policies and cases cited in Thaler v. Perlmutter reflect an outdated understanding of creativity and authorship in the age of generative AI. The court's reliance on precedents established in a different technological context calls for a reevaluation of copyright policies to accommodate the evolving nature of creative processes involving AI. A more nuanced and adaptive approach to copyright law is necessary to address the unique challenges posed by AI-generated art and foster innovation in the realm of intellectual property.

In countries like the UK, where judicial processes rely heavily on precedents, there's an even greater urgency for the introduction of new policies. The appealed case of Thaler v. Comptroller-General of Patents, Designs and TradeMarks in the UK, concerning the eligibility of AI-generated inventions for patent protection, similarly raises crucial questions about the applicability of existing policies and precedents. [10] In this instance, Dr. Stephen Thaler sought patents for inventions created by another AI system named DABUS, which he also owned. The UK Intellectual Property Office (UKIPO) rejected the applications, asserting that patents could only be granted to inventions created by humans.

Dr. Thaler again attempts to challenge the traditional understanding that patents are exclusively reserved for inventions originating from human inventors. The court references Yeda Research and Development Company Ltd v. Rhone-Poulenc Rorer International Holdings throughout its decision as it is a foundational case that redefines several aspects of patents but most notably defines what the “inventor” of a patent is. [11] The court emphasizes 'the inventive step' as crucial for patentable technologies, referring to the stage where a human conceives the idea. The court, once again, leans on precedents and definitions that were set in an era when the concept of AI autonomously creating patentable inventions was more theoretical than practical.

The original case bases a large part of its argument in the Patents Act of 1977, as the court states that “there is no suggestion that Dr. Thaler had any role in the creation of the technology at hand other than filing the statement of inventorship fully and honestly.” [12] The case notably cites several cases, including Jeffreys v. Boosey (1854) and Celgard LLC v. Shenzhen Senior Technology Material Co (2020), to demonstrate that “an invention is a piece of information” and “it would have alarming consequences if there was property in information.” [13] The courts use of cases dating back to the 1800s demonstrates the importance of precedent in the UK, highlighting again the need for new policies.

The court continues the discussion on information by citing the difference between contractual rights and property rights in the case of Franklin v. Giddins (1978). [14] In Franklin v. Giddins, the DNA of the plant was claimed to be confidential information protected through obligation of confidence. The court in bringing up this case invalidates the argument of ownership of the AI software. The court presents a compelling case that humans must have a role in the creation of AI works, but fails to recognize the collaborative role that humans played in the development of AI systems. In the context of AI-generated inventions, humans are involved in designing, programming, and overseeing the AI. Ignoring this collaborative dimension oversimplifies the dynamic relationship between humans and AI in the inventive process.

The policies and court decisions referenced in the UK case, though thorough, demonstrate an outdated understanding of invention and authorship in the era of advanced AI. The court's reliance on precedents established in a different technological landscape underscores the need for a reassessment of patent policies to accommodate the evolving nature of inventive processes involving AI. A more adaptive and nuanced approach to patent law is essential to address the distinctive challenges posed by AI-generated inventions and to encourage innovation in the field of intellectual property.

The legal cases of Thaler v. Perlmutter in the United States and the UK case Thaler v. Comptroller-General of Patents, Designs and TradeMarks have brought to light critical issues surrounding the eligibility of AI-generated creations for intellectual property protection. Central to both cases is the adherence to the requirement of human authorship or inventorship, an outdated notion deeply embedded in existing legal frameworks. Drawing upon outdated policies and precedents, such as the Burrow-Giles Lithographic Co. v. Sarony case in the US and the Yeda Research case in the UK, the courts find themselves grappling with the challenges posed by advanced AI capabilities that were not contemplated when these frameworks were established.

Both cases share a common concern – the failure to fully recognize the collaborative nature of AI creation. While Thaler's case in the US challenges the exclusive nature of copyright protection for human-authored works, the UK case questions the interpretation of inventorship in the context of AI-generated inventions. Despite humans playing pivotal roles in designing, training, and overseeing AI systems, the courts lean towards a strict interpretation, overlooking the intricate relationship between human guidance and AI autonomy.

The technological advancements in AI underscore the inadequacy of existing legal frameworks, calling for a reevaluation of intellectual property laws to accommodate the current capabilities of AI systems. The shared concerns regarding outdated legal frameworks and the failure to acknowledge collaboration emphasize the need for a more adaptive approach. In conclusion, both the US and the UK must develop new legal frameworks that consider the unique challenges posed by AI-generated creations, fostering innovation while ensuring that intellectual property laws evolve alongside the rapidly changing landscape of AI technology. Legislative updates and policy revisions, likely to be slow given the amount of precedent that surrounds property protection, are imperative to acknowledge the collaborative nature of AI-human partnerships in creative and inventive processes, thereby propelling intellectual property law into the future.

As AI becomes increasingly integral to diverse sectors of society, the legal landscape faces pressing challenges in adapting to the evolving nature of technology, particularly in the realm of patents. In this era marked by rapid AI innovation, a compelling imperative emerges – a call for a deliberate and adaptive legal framework that not only harnesses the potential of AI but also upholds the foundational principles of invention and creativity.

A pointed critique lies in the failure of existing policies to keep pace with the relentless growth of AI. The technological trajectory of AI is dynamic and evolving, presenting an ongoing challenge for legal systems rooted in case law. The inherent flexibility and adaptability of AI demand a corresponding responsiveness in legal frameworks, a characteristic that is conspicuously lacking in the current paradigms. [15]

Moreover, the insistence on strictly human-centric definitions of inventorship and authorship, as observed in both jurisdictions, underscores a resistance to recognizing the collaborative nature of AI-human partnerships. This resistance, while serving as protection against a system of unknown capabilities, not only impedes the progress of innovation, but also risks stifling the inherent potential of AI to contribute meaningfully to creative and inventive processes.

The prevailing legal frameworks in the United States and the UK reveal an urgent need for political and legislative introspection. A forward-looking approach is required, one that acknowledges the dynamic landscape of AI technology and addresses the persistent challenges that will continue to emerge. The call for legal evolution is not just a response to the current state of affairs; it is an acknowledgment that the trajectory of AI growth demands proactive and adaptive policies to ensure that intellectual property laws remain relevant and conducive to innovation. As the international community grapples with the complexities of AI, a reimagining of legal doctrines becomes paramount to foster an environment that nurtures both creativity and technological advancement.

 Edited by William Tang

[1] Haan, Kathy. “24 Top AI Statistics & Trends in 2023 – Forbes Advisor.” Edited by Rob Watts. www.forbes.com, April 25, 2023. https://www.forbes.com/advisor/business/ai-statistics/.

[2] Thaler v. Perlmutter, Case 1:22-cv-01564-BAH, 1 (D.D.C., 2023). https://www.copyright.gov/ai/docs/district-court-decision-affirming-refusal-of-registration.pdf

[3] Thaler v. Perlmutter, Case 1:22-cv-01564-BAH, 4 (D.D.C., 2023)

[4] Thaler v. Perlmutter, Case 1:22-cv-01564-BAH, 7-8 (D.D.C., 2023)

[5] Thaler v. Perlmutter, Case 1:22-cv-01564-BAH, 8 (D.D.C., 2023)

[6] Thaler v. Perlmutter, Case 1:22-cv-01564-BAH, 9 (D.D.C., 2023)

[7] Thaler v. Perlmutter, Case 1:22-cv-01564-BAH, 9 (D.D.C., 2023)

[8] Thaler v. Perlmutter, Case 1:22-cv-01564-BAH, 11 (D.D.C., 2023)

[9] Wilson, H. James, and Paul R. Daugherty. “Collaborative Intelligence: Humans and AI Are Joining Forces.” Harvard Business Review, July 2018. https://hbr.org/2018/07/collaborative-intelligence-humans-and-ai-are-joining-forces.

[10] Thaler v. Comptroller-General of Patents, Designs and TradeMarks, EWCA Civ 1374, 2-3 (UKSC 2023). https://www.supremecourt.uk/cases/docs/uksc-2021-0201-judgment.pdf

[11] Thaler v. Comptroller-General of Patents, Designs and TradeMarks, EWCA Civ 1374, 15 (UKSC 2023).

[12] Thaler v. Comptroller-General of Patents, Designs and TradeMarks, EWCA Civ 1374, 2-28 (RCOJ 2021).

[13] Thaler v. Comptroller-General of Patents, Designs and TradeMarks, EWCA Civ 1374, 2-28 (RCOJ 2021).

[14] Thaler v. Comptroller-General of Patents, Designs and TradeMarks, EWCA Civ 1374, 3 (UKSC 2023).

[15] Pal, Subharun. "The Legal Conundrum: Intellectual Property Rights in The Era of Artificial Intelligence." 2 (2023). https://www.researchgate.net/publication/371504157_The_Legal_Conundrum_Intellectual_Property_Rights_in_The_Era_of_Artificial_Intelligence