By Ali Ebshara, LL.M. 2021
On September 2, 2021, the U.S. District Court for the Eastern District of Virginia in Thaler v. Hirshfeld ruled in favor of the United States Patent and Trademark Office (USPTO), stating that an artificial intelligence (AI) cannot be listed as an inventor on a patent.1 The plaintiff alleged he had created an AI capable of forming patentable creations and filed for a patent application with the AI listed as an inventor.2 But the USPTO refused the application and stated that patent laws do not expand to include machines within the definition of an inventor.3
In coming to its decision, the court considered the interpretation of the language in the Patent Act.4 Specifically, the court considered the Act’s use of the term “individual” and whether it can expand to non-humans.5 The court relied on the Dictionary Act to find that the term “individual” should be interpreted through its common usage.6 The court also considered the use of personal pronouns like “himself” or “herself,” citing a case where the Supreme Court held that words in statutes should be interpreted by the words adjacent to them.7 In this sense, the pronouns used before the term “individual” help cement the fact that Congress had intended the term “individual” to refer to a person. The court also relied on the Federal Circuit’s decisions holding that being an inventor requires mental capacity to conceive the invention and that only natural persons can be inventors.8
The plaintiff had argued that the Patent Act should incorporate AI, as it would incentivize creation through AI and better protect moral rights. The plaintiff reasoned that human inventors’ moral rights currently stand unprotected due to the necessity of a human inventor to list themselves as the sole inventor, even when they have only had partial contribution alongside an AI.9 The court rejected this argument.10 In reliance on Fisons PLC v. Quigg, the court held that policy considerations cannot change the language of statutes and that such considerations should be left for Congress, not the courts.11 The court also emphasized the USPTO’s efforts in deciding whether an AI can be labeled as an inventor, referencing the USPTO’s research that found AI to be narrow and dependent on human intervention, and that AI able to be independently creative does not currently exist.12
In an almost directly contradicting decision, the Federal Court of Australia found that an AI can be listed as an inventor.13 The Australian court found that the ordinary usage of the term “inventor” is an agent noun that is not limited to a person but extends to anything that invents.14 Further, the court agreed with the plaintiff’s argument on the fact that listing AI as inventors would allow for better disclosure practices in patents and create greater financial incentives to create AI.15
Although the U.S. district court decision might seem like a negative outlook on the future of AI regulation within the patent space, it can be interpreted positively—the court simply indicated that AI cannot be listed as inventors yet. The court had acknowledged in its decision that a future where AI can be listed as inventors is likely, but we are simply not there yet.16 The Australian decision indicates that there is a movement towards acknowledging AI rights, and it is likely that the United States will follow suit within the near future.