Patentability Criteria for AI Inventions: Scope and Limitations

The recent Exeter Blockchain and Law Conference 2023 shed light on the growing challenges posed by artificial intelligence (AI) in the realm of patent law. Nigel Hanley, head of the examining group at the UK Intellectual Property Office (UKIPO), emphasized the pressing need for clarity in the treatment of AI within patent claims. With a dearth of legal precedents, the conference delved into the complexities of determining whether an AI can be recognized as an inventor.

The UK IPO’s evaluation guidelines for AI-related patent claims

Despite the evolving landscape, the UKIPO currently operates within defined parameters for evaluating patent claims involving AI. The office adheres to a working definition of AI as technology capable of emulating human intelligence to accomplish tasks such as visual perception, speech recognition, and language translation. Notably, the UKIPO maintains an unwavering stance, urging adherence to established positions concerning computer programs. Consequently, specific exclusions bar patent approval for discoveries, scientific theories, mathematical methods, and mere computer programs.

To surmount these exclusions, any AI-related invention must exhibit a “technical contribution” by solving problems beyond or within the computer or by presenting a new computer operation in a technical sense. In essence, the UKIPO emphasizes the necessity of a substantial technical element beyond the AI’s computational capabilities for patent approval.

Identifying Patentable AI Inventions

Hanley’s illustrative examples underscored the challenges of obtaining AI-related patents. Instances of AI trading assistants and AI-driven financial planners were deemed non-patentable due to their classification as methods of doing business, which fail to exhibit technical contributions. Notably, the emphasis on technical contributions within AI inventions reiterated the necessity of problem-solving capabilities beyond the AI’s computational framework.

In contrast, Hanley highlighted the potential patentability of AI that modifies existing data for enhanced training methodologies. This scenario exemplifies a technical contribution by introducing novel ways to operate computers, thus meeting the criteria for patent approval.

Contentious issue: Can AI serve as an inventor?

A crucial point of contention arose during the Q&A session, focusing on whether AI can be recognized as an inventor. Hanley’s response revealed that the UKIPO remains oblivious to patent applications drafted by AI, underlining the challenge of distinguishing human versus AI-generated submissions. However, the pivotal question remains unsettled: Can an AI be acknowledged as an inventor, or must the inventor be human?

The ongoing legal dispute in the U.K. Supreme Court, instigated by entrepreneur Dr. Stephen Thaler, centers on the patentability of AI-derived inventions and the potential inclusion of AI systems as inventors. While the UKIPO and the lower courts have maintained a position rejecting AI as inventors, the case’s implications extend to pivotal legal questions concerning the U.K. Patents Act 1977 and its application in the AI domain.

Considering the global context, countries like the U.S. and Australia have followed suit in rejecting AI patents, except for South Africa, which stands as the sole exception. Hanley refrained from providing an official stance on whether AI can serve as an inventor, acknowledging the pending judicial decision and its potential impact on the legal landscape.

Embracing the intersection of AI and blockchain

As the conference emphasized the intricate relationship between AI and legal frameworks, the intersection of artificial intelligence with blockchain technology emerged as a promising avenue for resolving existing challenges. Notably, the integration of blockchain in AI applications holds the potential to establish transparent and immutable records of AI-generated inventions, thereby addressing concerns of ownership and authenticity in the patenting process.

With the future of AI patents hinging on the upcoming judicial ruling, the integration of blockchain technology stands poised to revolutionize the patenting process, fostering an environment of transparency and accountability in the realm of AI-driven innovation.