AI and the Role of Inventorship in UK Patent Applications

UK Supreme Court rules that AI systems cannot be named as an inventor for UK patent applications

Introduction

Against a backdrop of rapid technological advancement and innovation, the growing sophistication of AI systems in recent years has raised complex questions of how the existing patent system will protect AI-generated inventions. In particular, consideration of the legal position regarding whether AI systems can be acknowledged as an "inventor" in the context of patent applications has sparked considerable discussion among courts across multiple jurisdictions (the "DABUS Case"). 

On 20th December 2023, in a landmark judgement the UK Supreme Court held that an inventor for the purposes of the Patents Act 1977 must be a natural person and confirmed that an AI system cannot be designated as an inventor to secure patent rights. However it should be noted that the UK Supreme Court decision does not concern the wider issue of whether inventions generated by AI systems are patentable under UK law. 

Although the highly anticipated decision brings an end to the DABUS Case in the UK for now, the significance of the judgement in the long-term is that it exposes an intrinsic divide between: i) existing patent laws designed to promote human ingenuity; and ii) the complexities of designating and assigning legal rights to 'inventive' AI systems, which have evolved in the past year at an unprecedented pace. 

In this article, we'll examine the background to the DABUS Case, consider the decision of the Supreme Court, and explore the potential practical and legal implications of the judgement.

Background

The case of Thaler v Comptroller General of Patents Trade Marks and Designs revolves around two patent applications filed by Dr. Stephen Thaler in relation to two different inventions reportedly created without any human involvement by the AI system named 'DABUS' (Device for Autonomous Bootstrapping of Unified Sentience). Developed by Dr. Thaler, the AI system leverages neural network technologies and generated two distinct inventions, being a fractal-based design for a food container and an emergency light system.  

In late 2018, Dr. Thaler submitted patent applications for these creations, crediting DABUS as the sole inventor, and stating that he was entitled to apply for the patents by virtue of his ownership of DABUS. However, the UK Intellectual Property Office ("UK IPO") rejected the applications in 2019 and subsequent appeals to both the UK High Court and Court of Appeal were dismissed. In summary, the grounds for refusal rested on two primary factors: 

1. UK patent law mandates the identification of a human inventor and as such DABUS did not qualify as an inventor under the 1977 Patent Act; and
2. no existing framework allows Dr. Thaler to claim patents derived from inventions independently generated by an AI system.

The Supreme Court Judgement

The Supreme Court judgement delivered by Lord Kitchin concluded the following:

• As a matter of statutory interpretation, an inventor within the meaning of the 1977 Patent Act must be a natural person. DABUS, being a machine entity, is not a natural person with legal personality, and as such falls outside the scope of the meaning of 'inventor' for the purposes of the 1977 Patent Act. 
• Since an AI system doesn't meet the qualification of an inventor, it cannot transfer the patent filing rights related to the invention. Dr. Thaler's assertion of ownership through the "doctrine of accession," asserting rights due to his ownership of DABUS, was deemed erroneous. Consequently, this claim did not grant him the authority to seek or secure patents for the machine-generated inventions by DABUS.
• The IPO had the authority to consider Dr. Thaler's applications as withdrawn due to his non-compliance with section 13(2) of the Patents Act 1977. In this case, Dr. Thaler failed to specify any individual believed to be the inventor, and his claim of entitlement to apply for the patent was based on the misguided position that DABUS held the status of an inventor.

While drawing these conclusions, the Court explicitly emphasized that the case didn't address the wider debate surrounding the patentability of technical advancements originating from AI or the potential expansion of the term "inventor" to encompass AI entities. Notably however, the court acknowledged that the debate highlights long-term questions around patentability in light of rapidly developing technological improvements, and that such issues pose potential policy considerations concerning the purpose and design of the existing patent system. In respect of this matter, the Court clarified that its role was confined to interpreting existing laws rather than engaging in the broader discourse on the purpose and scope of patents in relation to AI-generated innovations.

Analysis and Implications

The judgement itself means that the UK is now aligned with the legal position adopted by courts in the European Union and the United States in respect of the DABUS case, and importantly provides legal clarity for businesses in relation to patentability of creations by AI technologies. 

For the immediate future, this outcome is not expected to significantly disrupt businesses. Presently, AI systems predominantly serve as tools to human inventors, so businesses won't encounter difficulty in identifying one or more individuals to designate as inventors in their patent submissions. Additionally, businesses can find comfort in the understanding that challenges to their eligibility to file for patents, according to the Patents Act 1977, can only arise from individuals claiming their entitlement to apply for the patent—not from an AI system.

However, as noted above, the Supreme Court's verdict did not deal with the broader question of whether AI-generated outputs themselves are patentable, and potentially gives rise to future discussions among legislators regarding whether the definition of "inventor" should be expanded to cover AI. In its response to the judgement, the IPO stated that it recognises "that there are legitimate questions as to how the patent system and indeed intellectual property more broadly should handle [AI] creations". The increasing capability of AI to autonomously generate inventions is likely to be enhanced given the current pace of technological advancements within the field. From a long-term perspective, this may potentially put pressure on policymakers to address AI innovation and impose legislative interventions to facilitate patentability of inventions which are independently created by AI systems. 

In June 2022, the IPO had concluded that there was no evidence that UK patent law is currently inappropriate to protect inventions made using AI. However, since this time period, there have been significant developments in AI models. Over the past year alone, we have seen the evolution of generative AI to match and outperform human output, and the increasing impact of AI across multiple fields, including medicine and science. These advancements represent a fundamental shift in the legal landscape in respect of AI inventorship and patent law. Consequently, as the UK aims to position itself as a major AI leader per Prime Minister Rishi Sunak's policy objectives, it seems that legislative change is inevitable in order to incentivise innovation. In the meantime, businesses that own 'inventive' AI systems should maintain sufficient human participation in the inventive process to adhere to patent application conditions, and explore alternative means of safeguarding innovations generated by AI.

Conclusion

Although the judgement and fallout of the DABUS case has provided a definitive position on the current law in the UK, the case has reflected an inherent tension between existing patent law and the inevitability of technological innovation, particularly in respect of the evolving nature and use cases of AI. The judgement may bring legal clarity for now, but the case itself raises broader questions on the nature of the existing legal framework, and whether it is, or even should be, sufficiently adaptable and flexible to accommodate non-human inventors in the future.

Authors: Sunny Kumar, Partner; Nilesh Ray, Junior Associate