In July 2019, Thaler sought patent protection for two of DABUS’ putative inventions by filing two patent applications with the U.S. Patent and Trademark Office (PTO), listing DABUS as the sole inventor on both applications. In lieu of an inventor’s last name, Thaler wrote on the applications that “the invention [was] generated by artificial intelligence.” The PTO concluded that both applications lacked a valid inventor and were, therefore, incomplete. It sent Thaler a Notice to File Missing Parts of Nonprovisional Application for each application and requested that Thaler identify valid inventors. Thaler petitioned the PTO to vacate the Notice, which the PTO denied. Thaler sought reconsideration of the denial (also denied), and then pursued judicial review of the PTO’s decisions in district court. The district court granted summary judgment to the PTO, concluding that an “inventor” under the U.S. Patent Act must be an “individual” and the plain meaning of “individual” as used in the statute is a natural person.
Thus, the question before the Federal Circuit in Thaler was whether anyone other than a human being can be an inventor of a U.S. patent. In examining this question, the Federal Circuit looked to the statutory language of the U.S. Patent Act, which defines an “inventor” of a patent as “the individual…who invented or discovered the subject matter of the invention.” However, as the Federal Circuit observed, the Patent Act does not define “individual.”
Thus, to interpret “individual” in the Patent Act, the Federal Circuit applied various principles of statutory interpretation and looked to (1) U.S. Supreme Court precedent interpreting “individual” when used in statutes to mean a human being, unless Congress intended to deviate from the default meaning, (2) the ordinary meaning of “individual” as a human being, (3) the Patent Act itself, which uses personal pronouns such as “himself” or “herself” to refer to an “individual” (not “itself”) and requires inventors to submit an oath or declaration that they are the inventor, and (4) the Federal Circuit’s own precedent that inventors must be human beings, not corporations or sovereigns.
The Federal Circuit concluded that the U.S. Patent Act unambiguously “requires that inventors must be natural persons; that is, human beings.”
This ruling adds to a growing body of decisions around the world finding that AI may not be an inventor. However, the decision in Thaler also reinforces that whether AI may be an inventor turns on interpretation of the particular statute at issue. For example, among Thaler’s many arguments to the Federal Circuit was that South Africa has granted patents with DABUS as an inventor. The Federal Circuit rejected this as a reason to recognize AI as an inventor of a U.S. patent, reasoning that “[t]his foreign patent office was not interpreting our Patent Act,” and thus “[i]ts determination does not alter our conclusion.”
Whether patent laws will evolve to recognize AI as an inventor remains to be seen. Thaler and his supporters have argued that the laws must evolve in this fashion in order to encourage innovation, and have applied for patent protection naming DABUS as the inventor in a variety of jurisdictions. While the South African IP office allowed DABUS as an inventor, and a court in Australia concluded in a later overruled opinion that inventions conceived by AI systems must not be excluded from patent protection, these views appear to be the exception rather than the rule. As it stands now, AI is not recognized as an inventor under the patent laws of the United States, Europe, and Australia.
And as to Thaler’s argument that patent law must further develop in order to foster innovation, it is worth noting that many commentators question whether AI systems are even capable of independently conceiving inventions. According to the majority view, they cannot, and the theory that AI can make inventions should be delegated to the realm of science fiction. While the technical and philosophical debate continues, this does not mean that owners of AI systems are bereft of protection. If a human uses AI to develop a solution or to verify an outcome, then this human can be named as the inventor. What’s more, AI-related technologies may be protected as computer-implemented inventions. Thus, the primary question for patent applicants is not whether, but how to apply for patent protection of AI-related inventions. For the time being, the answer to that question in the United States is that the applicant can only identify a human being – not AI – as the inventor.
Client Alert 2022-213
