The Cyberlaw Clinic recently submitted a comment in response to a request from the United States Patent and Trademark Office about the patentability of artificial intelligence (“AI”) inventions. The comment was submitted on behalf of the R Street Institute, a nonprofit, nonpartisan, public policy research organization. The USPTO sought guidance on handling patent applications for inventions created by generative AI technology. Addressing some of the many questions posed by the USPTO, the comment discusses whether entities other than natural persons should be able to own a patent, the impact that AI technology has on the level of a person of ordinary skill in the art, and whether patent laws need to be revised to account for AI inventions.
The issue of patentability of inventions created by generative AI recently increased in relevance, because the first-ever patent applications listing a computer as the inventor were filed in patent offices around the world. Inventors have historically always been human beings, and the USPTO has yet to issue guidance about patentability considerations for inventions that are generated either wholly or partially by computers.
The comment outlines both current positive law as well as policy considerations in response to the questions posed by the USPTO. Among the issues presented, the comment argues that computers cannot be inventors under current law because there is a statutory requirement that inventors must be individuals, defined as “natural persons,” and because inventorship requires conception of the idea in the mind of the inventor, and computers do not have minds. Additionally, the comment discusses which natural persons may have a claim for inventorship of inventions created by AI.
The comment also notes that the obviousness inquiry for patentability may need to evolve. Computers have faster computing power than humans and access to expansive data sets, so what is difficult for a human to create may well be “obvious” to a computer or to an inventor with access to AI technology. The availability of AI technology to innovators in a particular field thus impacts which inventions are deemed obvious. Finally, the comment addresses the potential benefits of keeping AI-generated inventions in the public domain, because patenting these inventions could generate costly litigation about inventorship, lead to unproductive patent races, and disincentivize human innovation.
The Clinic worked closely with Charles Duan of the R Street Institute on the comment.