The UK Supreme Court unanimously ruled Wednesday that an “inventor” under the Patents Act 1977  must be a “natural person” and not artificial intelligence (AI). The Supreme Court heard the landmark appeal, Thaler v Comptroller-General of Patents, Designs and Trade Marks [2023], after it was initially dismissed in the High Court and Court of Appeal in 2021.

The appellant, Dr. Stephen Thaler, filed two applications for patents in 2018 under the 1977 Act, with forms that stated that the inventor was a machine called DABUS, which was “acting autonomously and powered by AI.” Thaler maintained that he had the right to be granted patents due to his ownership of DABUS. The Comptroller-General of Patents refused to grant the patents, and the appellant brought the issue to the High Court, and then the Court of Appeal, where the matter was dismissed.

The appeal dealt with two main issues: the question of whether AI could be an inventor under the 1977 Act and whether the appellant was entitled to apply for patents as the owner of DABUS. The court used case law authorities to interpret the meaning and scope of an “inventor,” and found that the only reasonable interpretation is that the inventor must be a “natural person” “who came up with the inventive concept.” This therefore cannot extend to machines and AI. The court noted that when crafting the legislation, Parliament did not anticipate that machines would be acting as inventors.

Regarding the second issue, the court rejected the appellant’s argument that his ownership of DABUS resulted in a de facto right to grant of patents. Thaler’s argument hinged on the “doctrine of accession.” However, the court rejected this and said that the doctrine of accession could not apply here, as DABUS was not the inventor and an invention cannot be “passed” to the owner of the machine that created it.

The landmark decision is a significant development in the AI landscape, following new developments in AI regulations in the EU’s AI Act and the world’s first AI Safety Summit which took place in the UK in November. In early December 2023, the EU reached a provisional agreement for the AI Act, the world’s first artificial intelligence regulation. The legislation seeks to “develop an ecosystem of trust by proposing a framework for trustworthy AI”.