It was a court case that many in the creative and technology industries had been watching closely, and on November 4, 2025, the British High Court delivered its long-awaited verdict: Stability AI, the developer behind the popular Stable Diffusion image generator, had largely prevailed against Getty Images in a landmark intellectual property lawsuit. Yet, as the dust settles, both sides are left claiming victory, while experts say some of the most pressing questions about artificial intelligence and copyright remain unresolved.
The case began in January 2023 when Seattle-based Getty Images, a global powerhouse in stock photography, filed suit in the UK High Court. Getty accused Stability AI of scraping millions—reportedly 12 million—of its copyrighted images, along with associated metadata, to train Stable Diffusion, an AI model capable of generating realistic images from text prompts. Getty’s allegations didn’t stop at copyright infringement; the company also claimed that AI-generated outputs sometimes reproduced Getty’s content, complete with its watermark, and that this amounted to trademark infringement, database rights infringement, and even “passing off”—the suggestion that AI-generated images were endorsed by or created by Getty itself.
Stability AI, for its part, mounted a robust defense. The company argued that the model’s training took place outside the UK—on computers run by U.S. tech giant Amazon—and that any similarities in output were due to users’ prompts, not the company’s deliberate actions. Stability also pushed back on the trademark claims, insisting there was no “in the course of trade” context that would make them liable.
As the trial unfolded in June 2025, a pivotal moment arrived: Getty dropped its primary copyright infringement claims, citing evidential and jurisdictional difficulties. According to reporting by the Associated Press, this move signaled that Getty did not think its arguments about the legality of AI model training under UK law would succeed. That left the court to consider narrower claims—trademark infringement (due to the appearance of Getty’s watermark in AI outputs), secondary copyright infringement, database rights, and passing off.
When the verdict finally arrived, it was something of a split decision. High Court judge Mrs Justice Joanna Smith ruled that Getty had succeeded, in part, on its trademark claim. The court found that the inclusion of Getty’s watermarks in some AI-generated images did indeed infringe on Getty’s trademark rights. However, the judge described her findings as “historic and extremely limited in scope,” and emphasized that she had not found evidence of widespread infringement. “While I have found instances of trademark infringement, I have been unable to determine that these were widespread,” she stated, as reported by the Baltimore Sun.
On the more consequential copyright front, Stability AI emerged victorious. The court dismissed Getty’s secondary copyright infringement claim, and since the broader copyright-training claim had already been abandoned, it was never adjudicated. Justice Smith wrote that Stable Diffusion’s AI “does not store or reproduce any Copyright Works (and has never done so),” a finding that shielded Stability from liability under UK copyright law.
Both companies tried to put the best spin on the outcome. Getty Images issued a statement declaring, “This is a significant win for intellectual property owners,” highlighting the court’s confirmation that responsibility for trademark infringement lies with the model provider, not the end user. Stability’s general counsel, Christian Dowell, responded with satisfaction: “This final ruling ultimately resolves the copyright concerns that were the core issue.”
Yet, for all the rhetoric, legal experts were quick to point out the limitations of the ruling. Simon Barker, Partner and Head of Intellectual Property at Freeths, told reporters that the decision “draws a line to say that training an AI model on copyright works, without storing or reproducing those works in the model itself, does not amount to secondary copyright infringement under UK law.” Barker suggested that AI developers could take some comfort from the case, knowing that the mere act of training on large datasets is not, by itself, a copyright violation in the UK. However, he warned that if AI-generated outputs reproduce protected trademarks, such as watermarks, in a way that could confuse consumers, companies still risk liability.
Iain Connor, an intellectual property partner at Michelmores, was even more blunt. “The most significant AI case to reach the English High Court has been decided and has turned out to be a massive damp squib,” he said. Connor noted that because Getty dropped its main copyright claims on jurisdictional grounds, “the judge had no opportunity to rule in general terms on the lawfulness of AI’s use of copyright-protected ‘input materials’ and whether an AI model’s ‘output’ infringed such copyrights.” In other words, the legal community is still waiting for a definitive answer on the thorniest issues at the intersection of AI and copyright.
The ruling comes amid a broader wave of litigation as creative industries push back against the use of their works to train generative AI systems. In the United States, Getty is pursuing a separate copyright infringement lawsuit against Stability AI, having refiled its case in a San Francisco federal court in August 2025. Meanwhile, other tech companies are facing similar challenges: Anthropic agreed to pay $1.5 billion to settle a class-action lawsuit by authors over training its Claude chatbot on pirated works, while a federal judge dismissed a comparable suit against Meta Platforms. Media giants like Warner Bros., Disney, and Universal have also launched lawsuits against AI startups accused of generating unauthorized images of their copyrighted characters.
Financial markets responded in real time to the drama. Getty’s shares dipped 3% before the U.S. market opened on November 4, 2025, following the court’s verdict, though they had risen a few days earlier after the company announced an AI licensing deal. The mixed result reflects the uncertainty that still hangs over the sector.
Ultimately, the High Court’s decision provides some clarity for now, but leaves many crucial questions unanswered—especially around the legality of training AI models on copyrighted materials, the responsibilities of model providers versus users, and the global jurisdiction of such disputes. As Justice Smith herself observed, there is “very real societal importance” in deciding how to balance the interests of creative industries and technological innovators. For now, though, the legal battle between Getty Images and Stability AI ends with a muted resolution: a narrow trademark victory for Getty, a broader win for Stability AI, and a legal landscape that remains, in many ways, as unsettled as ever.
