UK Supreme Court rules AI cannot be a patent inventor

The UK Supreme Court unanimously dismissed Stephen Thaler’s appeal in Thaler v Comptroller-General of Patents, Designs and Trade Marks, ruling that an “invention” under section 7 of the Patents Act 1977 must have a human inventor. Thaler had filed two UK patent applications in 2018 — for a food container and a flashing light — naming his AI system DABUS (“Device for the Autonomous Bootstrapping of Unified Sentience”) as sole inventor, with himself listed as applicant by virtue of owning the machine.

The five justices, led by Lord Hodge, held that the statute’s use of “inventor” required a natural person and that Thaler could not derive a right to apply for a patent under section 13 without identifying one. Because no human inventor was named, the applications were treated as having been withdrawn. The court was explicit that it was not ruling on the broader policy question of whether inventions autonomously generated by AI systems should be patentable at all — only on how the existing 1977 Act, drafted decades before generative AI, allocates inventorship under its current wording.

The ruling followed the same reasoning as the US Federal Circuit’s 2022 decision in Thaler v. Vidal, which read the US Patent Act’s reference to an “individual” the same way, and Thaler had lost equivalent challenges in other jurisdictions including the European Patent Office and Australia. The decision left intact the practical route open to anyone with an AI-assisted invention: name the human researcher who directed, selected or interpreted the AI’s output as the legal inventor, a workaround the court did not need to rule on but that patent practitioners generally regarded as unaffected by the judgment.