UK Court Sides with Stability AI in Getty Images Copyright Case, But Flags Trademark Violation

A U.K. High Court has largely ruled in favor of Stability AI Ltd. in a landmark copyright case brought by Getty Images Holdings Inc., marking a significant development in how courts interpret AI training and copyright law.

Getty sued the London-based startup — creator of the Stable Diffusion text-to-image model — in early 2023, alleging that millions of its copyrighted photos were unlawfully used to train the AI system.

However, Justice Joanna Smith dismissed the copyright infringement claims, ruling that the AI models were trained outside the U.K.

“Getty Images may be able to maintain such a case in the jurisdiction where the Model was in fact trained, but there is no basis for that case in this jurisdiction,” Smith wrote.

The court also rejected Getty’s argument that Stable Diffusion’s training process produced infringing copies, finding that “the model weights are not themselves an infringing copy and they do not store an infringing copy.”

However, the court sided with Getty on trademark infringement, noting that some AI-generated images reproduced iStock watermarks, amounting to trademark misuse.

Stability AI welcomed the ruling, calling it a resolution to the “core copyright concerns.” Getty said it would pursue similar claims in the U.S., where another case is pending.

Following the decision, Getty’s shares dropped nearly 9%.

Recently, Stability AI has adopted a more collaborative and industry-aligned approach to developing AI models, announcing partnerships with Electronic Arts and Universal Music Group.

Earlier this year, AI startup Anthropic settled a major class-action lawsuit filed by U.S. authors over alleged copyright infringement involving 7 million pirated books. Facing potential damages of up to $900 billion, Anthropic called the case a possible “death knell.” Settlement terms remain confidential pending court approval.