Can Generative AI Be An Inventor?

Generative AI has evolved significantly through the integration of machine learning algorithms, allowing systems to create content, models, and solutions autonomously based on learned patterns and data. Can AI be an inventor? The Federal Circuit said no in its decision Thaler v. Vidal , denying a researcher’s claim of AI as the inventor of two patents. The court rejected the researcher’s interpretation of patent law, but it did not discuss the rights of AI or the nature of the invention. It reasoned that adjudicating on the patents’ inventorship only requires reading the language of the Patent Act, which explicitly requires inventors to be “individuals.” “When a statute unambiguously and directly answers the question before us, our analysis does not stray beyond the plain text,” the Federal Circuit held.

After the Federal Circuit denied treating AI as an inventor on its own, the question remains whether it can be named a co-inventor, and to what extent a human inventor can use AI in the process of invention. In response, in his executive order issued on October 30, 2023, President Biden asked the United States Patent and Trademark Office (USPTO) to draft guidelines on determining who the inventor is when an invention is developed using AI. The executive order instructed the USPTO to provide examples showcasing the various roles AI might play in the process of invention and to elucidate the evaluation process for determining inventorship in each specific role.

The question of AI inventorship gives rise to hard policy and ethical issues. Policy-wise, a disqualification akin to the Federal Circuit’s ruling may deter researchers and inventors from investing resources and time in developing generative AI due to the inability to safeguard resulting inventions. Content creators may encounter a substantial amount of freely available content that has the potential to weaken their position in the market. Ethical implication appears at the same time. Dr. Thaler, the creator of the generative AI system DABUS, claimed that not recognizing AI’s intellectual production by refusing to recognize its status as an inventor speciesism. In fact, the ethical inquiries about the societal status of AI surfaced long before its ascent. Can it be recognized as an independent creator or thinking being? How to better protect investment in generative AI? The USPTO is set to release its guidelines on February 23, 2024, and inventors and legal experts anticipate that they would address these pressing questions.