Ben Kovach


As artificial intelligence (AI) system’s capabilities advance, the law has struggled to keep pace. Nowhere is this more evident than patent law’s refusal to recognize AI as an inventor. This is precisely what happened when, in 2020, the U.S. Patent and Trademark Office (USPTO) ruled that it will not accept an AI system as a named inventor on a patent.

This note explores untenable legal fiction that the USPTO’s ruling has created. First, it explores the current state of AI systems, focusing on those capable of invention. Next, it examines patent law’s inventorship doctrine and the USPTO’s application of that doctrine to AI inventors. The note then explains that disallowing AI systems as inventors does not map well onto patent law’s most common justifications. Finally, the note recommends a solution that maximizes patent law’s incentive structure; AI systems should be allowed as named inventors when patent ownership has been pre-contracted away to a natural person. If patent ownership has not been pre-contracted, the idea should enter the public domain and be unpatentable.