Getty’s AI fight falls flat as London court delivers mixed verdict in landmark copyright case

LONDON : The court’s limited ruling, delivered Tuesday, has left unresolved questions about whether training AI models on copyrighted works constitutes infringement under UK law.

Seattle-based Getty Images, known for its extensive catalogue of editorial and creative stock content, had accused Stability AI of unlawfully using its images to train the company’s AI image generator, Stable Diffusion. The case was initially seen as a potential turning point for copyright enforcement in the digital era.

However, midway through the trial, Getty dropped its primary copyright claims, citing insufficient evidence about where Stable Diffusion was trained. Intellectual property lawyers later described the ruling as a “damp squib” for AI law, noting that key copyright questions remain unanswered.

Justice Joanna Smith ruled that Getty had succeeded “in part” on its trademark infringement claims, specifically over the unauthorized appearance of Getty’s watermark in AI-generated images. Yet she emphasized that her findings were “historic and extremely limited in scope.” The judge dismissed Getty’s broader claim of secondary copyright infringement, concluding that Stable Diffusion “does not store or reproduce any copyright works.”

The verdict underscored weaknesses in the UK’s existing copyright framework for emerging technologies. Stability AI welcomed the decision, with its general counsel, Christian Dowell, stating that it “ultimately resolves the copyright concerns that were the core issue.”

Getty, meanwhile, maintained that the judgment confirmed Stable Diffusion had infringed on its trademarks, calling it “a significant win for intellectual property owners.” The company also said the ruling established “a powerful precedent” that AI models could be subject to infringement claims, a principle it plans to pursue in its ongoing U.S. case against Stability AI.

Still, Getty’s shares fell 8.7% in premarket trading following the ruling. In a statement, the company urged governments to introduce stronger transparency rules to help creators protect their work without resorting to costly litigation.

Legal experts agreed the outcome left critical questions unanswered. Gill Dennis of Pinsent Masons said the government must now issue “clear, timely policy guidance” on whether training AI models on copyrighted materials constitutes infringement. Iain Connor of Michelmores added that the decision “leaves the UK without a meaningful verdict on the lawfulness of an AI model’s learning process.”

The case, which began in June with widespread attention for its potential to reshape intellectual property law, now concludes with limited legal clarity but significant implications.