AI's Impact on Patent Law: Challenges and Solutions

Since the inception of "artificial intelligence" (AI) by John McCarthy in 1956, the realm of AI has gone through periods of inflated expectations and gradual progress. However, recent breakthroughs, particularly exemplified by OpenAI's ChatGPT, have propelled AI into a practical sphere, resulting in the development of innovative products and services across various industries. This article delves into the significant challenges AI poses to the patent system and how it has sparked a global discussion on AI inventorship, nonobviousness, written descriptions, and enablement.

AI and Its Impact on Patent Law

The traditional patent laws were formulated in an era when inventions were predominantly the result of manual creative processes. These laws aimed to incentivize human inventors to share their creations with the public. With the rise of AI, which can automate parts of the inventive process, a series of complex questions emerge concerning how patentability standards should be applied to AI-assisted inventions.

For instance, Stephen Thaler's AI system, DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), has been designated as the sole inventor on global patents, triggering debates within patent offices and courts regarding AI's eligibility as an inventor. While Thaler's attempts have not succeeded so far, they have ignited substantial discourse on AI inventorship.

The United States Patent and Trademark Office (USPTO) issued a formal request for public input on AI inventorship and patentability. Questions were raised regarding the patentability of inventions created with AI assistance, the ownership implications of AI inventorship, the need for expanded guidance on AI inventorship, and the possibility of statutory changes to address AI's role in innovation. The USPTO is expected to publish responses to these questions, and regular stakeholder meetings on AI-related topics have been held.

AI as a Collaborative Tool

AI's primary contribution to invention lies in its ability to generate, evaluate, and filter potential innovations, such as in drug discovery. AI does not replace human inventors but augments their capabilities, leading to a collaborative human-computer approach that can expedite the creation of superior inventions. This parallels previous invention-facilitation tools like computer-aided design (CAD), 3D printing, and mathematical concepts.

The Nonobviousness Dilemma

If a human inventor utilizing AI can produce inventions more efficiently than one without AI, it raises questions about the nonobviousness requirement in patent law. In fields like drug development, where inventors often hold Ph.Ds, patent law recognizes a higher nonobviousness bar. Should the widespread use of AI in a specific domain warrant an elevated nonobviousness standard for patent claims? While no clear legal precedent exists, it's conceivable that patent examiners or defendants in infringement cases may soon advocate for a heightened standard due to AI's role in the creative process.

The Inscrutability of AI

Many AI systems produce outputs that are inscrutable to humans, making it challenging to satisfy patent law's written description and enablement requirements when drafting AI-related patent applications. However, this obstacle is not insurmountable, as seen in the chemical and biological fields, where inventors patent materials or processes without understanding the mechanisms fully. Techniques like product-by-process claims and method of use claims may also apply to AI-related patents, requiring patent professionals to adapt practices from diverse domains.

Conclusion:

The emergence of large language models has generated immense interest, yet their vast potential in automating invention remains underestimated by the general public and legal professionals. Nevertheless, the USPTO has recognized this trend and is actively seeking public input on adapting patent law to accommodate AI-driven innovation.

AI presents challenging but surmountable questions for patent law standards concerning inventorship, nonobviousness, written descriptions, and enablement. Addressing these issues proactively rather than reactively is crucial to ensuring that the patent system continues to foster innovation, considering the evolving landscape of invention powered by AI.