Summary
The Federal Circuit’s binding opinion affirming that AI systems cannot be named as inventors on U.S. patents. Stephen Thaler filed two patent applications listing his AI system DABUS as the sole inventor. The PTO rejected them; the district court agreed; the Federal Circuit affirmed, holding that the Patent Act’s use of “individual” unambiguously refers to natural persons (human beings). The opinion explicitly declines to decide whether AI-assisted inventions by humans are patentable — only that AI-only inventorship is not.
Key Points
- The holding: “Congress has determined that only a natural person can be an inventor, so AI cannot be.” The Patent Act defines “inventor” as an “individual,” and that word — consistently interpreted by SCOTUS — means a human being.
- The statutory analysis: The Patent Act uses “himself or herself” (not “itself”) when referring to inventors; it requires inventors to submit a sworn oath, which Thaler had to sign on DABUS’s behalf; Congress would have used “whoever” (a broader term) if it intended to include non-humans.
- What was NOT decided: The court explicitly carves out “the question of whether inventions made by human beings with the assistance of AI are eligible for patent protection.” That question remains open.
- Thaler’s arguments rejected: (1) “whoever” in §101 doesn’t override the §100(f) definition of “inventor”; (2) §103 (non-obviousness) is about how inventions are made, not who made them; (3) Constitutional avoidance canon doesn’t apply when the text is unambiguous.
- The global split: South Africa had granted DABUS a patent (applying its own law, not the U.S. Patent Act). Australia’s Federal Court initially granted DABUS a patent, later reversed. The global legal system is fragmented on AI inventorship.
- DABUS: “Device for the Autonomous Bootstrapping of Unified Science” — Thaler describes it as a “Creativity Machine” (connectionist AI). He filed identical applications in the U.S., UK, Europe, and Australia.
Newsletter Angles
- AI Legal Personhood: Thaler v. Vidal is the most legally authoritative U.S. ruling on AI personhood. It resolves the inventor question via plain statutory text — but it also points at the harder question: as AI-assisted invention becomes ubiquitous, who owns the output? The court’s carve-out (“human assistance” cases) will be the next frontier.
- The ownership gap: Under current law, if an AI invents something with no meaningful human contribution, the invention may be unpatentable (no human inventor) but also unprotectable by any other IP regime. This creates an incentive to lie about the degree of human contribution — the “human in the loop” will be legally required whether or not they’re intellectually present.
- Patent law as AI governance: This case is a preview of how existing IP frameworks will be stress-tested by AI. The answer “Congress can fix this” is technically true but politically unlikely — leaving judges to interpret century-old statutes against a technology they don’t understand.
Entities Mentioned
- DABUS — the AI system named as inventor; Thaler’s “Creativity Machine”; the subject of litigation across multiple jurisdictions
Concepts Mentioned
- AI Legal Personhood — Thaler v. Vidal is the definitive U.S. legal authority on AI inventor status; the case IS the concept in American law
- Tech-State Conflict — the state (PTO, courts) refusing to adapt IP law to accommodate AI-generated output
Quotes
“When a statute unambiguously and directly answers the question before us, our analysis does not stray beyond the plain text. Here, Congress has determined that only a natural person can be an inventor, so AI cannot be.”
“We are not confronted today with the question of whether inventions made by human beings with the assistance of AI are eligible for patent protection.” [The question left open]
Notes
This is the full Federal Circuit opinion — the primary legal source, not a summary. It was affirmed without SCOTUS review and remains controlling law in the U.S. The UK Supreme Court reached the same result in Thaler v. Comptroller-General (2023).