June 7, 2022

Making the Case for AI Inventorship: Thaler v. Vidal, Case No. 21-2347 (Fed. Cir.)

Holland & Knight IP/Decode Blog
Mark Masutani | Jacob W. S. Schneider
IP/Decode Blog

This post explores recent oral arguments in Thaler v. Vidal, Case No. 21-2347 (Fed. Cir.), the much-discussed patent case concerning AI inventorship. The case is a companion to Thaler's copyright case, which argues that AI can be an author of expressive works.

The case stems from the U.S. Patent and Trademark Office's (USPTO) refusal of two patent applications, both of which list an AI system as the sole inventor. Dr. Stephen Thaler argues that his AI system, the Device for the Autonomous Bootstrapping of Unified Sentience (DABUS), "invented" a flickering light that mimics neural activity and a fractal drink container that makes it easier for robots to grip it. The USPTO, however, determined that the Patent Act "limit[s] inventorship to natural persons." To be clear, Thaler is not seeking for DABUS to own the patents – Thaler claims ownership – but he wishes to give credit to DABUS as an inventor.

The case has made its way through an unsuccessful appeal to the Director of the USPTO, as well as the U.S. District Court for the Eastern District of Virginia. In both cases, Thaler's request was denied. Thaler appealed those decisions to the U.S. Court of Appeals for the Federal Circuit. The Court heard oral arguments on June 6, 2022, with Professor Ryan Abbott representing Thaler and Assistant U.S. Attorney Dennis Barghaan representing the USPTO.

Abbott leads an organization, The Artificial Inventor Project, dedicated to the procurement of intellectual property rights for AI-generated inventions, and his primary argument was grounded in policy. Abbott urged the court to consider the purpose of the Patent Act: to promote innovation. He pointed to language in the Patent Act that says "[p]atentability shall not be negated by the manner in which the invention was made," as evidence of Congress's indifference as to who or what is the inventor. According to Abbott, this language should be interpreted broadly to favor innovation, similar to how other language in the Patent Act has in the past. Moreover, he argued that we are moving into a new research paradigm where drug manufacturers are turning to AI systems for research and design in lieu of teams of pharmacologists. To not reward the drug manufacturers' investment with patent protection would be to discourage innovation.

Better Defined

Barghaan primarily argued for a straightforward application of statutory construction. The explicit definition of an "inventor" being an "individual" or "individuals" was added to the Patent Act less than 10 years ago, and the notion of an "individual" being a natural person is bolstered by the use of gendered pronouns in other places in the Patent Act. Barghaan stated that this plain meaning analysis ends the exercise despite any preferred policy outcome, laudable or not. He also made the case that Congress would be the appropriate body to address the panoply of complex questions (e.g., how could DABUS be deposed in an inventorship dispute?) that would arise from Abbott's preferred statutory construction.

The Court posed an question to Barghaan, but one slightly beyond of the facts of the instant case: Could the programmer of an AI system claim inventorship over an invention produced by the AI system if the programmer conceived of the invention when writing the code? Maybe, said Barghaan. Abbott agreed, while also warning of how AI systems are built in practice: Thousands of programmers might contribute to the development of an AI system while not appreciating what that system might eventually invent.

In Conclusion

The proceedings concluded with a moment of levity. Referring to his earlier point, Abbott stated, "[Innovation can come] from a person, a robot, or a room full of monkeys … all we really want is for the system to produce innovation." Abbott was quick to assert that they were not actually alleging that monkeys should be inventors and to forget that analogy, but the court was equally quick to prod whether Abbott was familiar with the Infinite Monkey Theorem and whether monkeys should be afforded copyright protection if they typed out complete works of Shakespeare (or Charles Dickens). Thinking on his toes, Abbott noted he was out of time.

While this case is too uncertain to accurately predict an outcome, based on Thaler's track record in earlier proceedings, he has a low chance of success. Thaler and the Artificial Inventor Project have filed numerous counterpart patent applications in foreign jurisdictions. Some jurisdictions – including Australia, the European Union, Germany and the United Kingdom – have already refused the applications on similar grounds to the USPTO. Only South Africa's patent office has granted an application, but the country does not have a substantive examination system.

Related Insights