Getty Images’ decision to drop its core copyright claim against Stability AI marks a turning point in how UK law approaches generative AI. The case reveals the limits of jurisdiction, the evolving nature of output liability and the urgent need for policy clarity. For AI developers and rights holders alike, it’s a wake-up call: enforcement must evolve as fast as the technology.

Published  3 July 2025

In a significant turn of events in one of the UK’s most closely watched AI-related legal battles, Getty Images has dropped its primary copyright infringement claim against Stability AI.

The case, which commenced in the High Court on 9 June 2025, centred on allegations that Stability unlawfully used millions of Getty’s photographs to train its image-generating model, Stable Diffusion.

Why the claim was dropped

Getty’s decision, announced during closing arguments on 26 June 2025, was described as a “pragmatic” move. The company acknowledged the evidential and jurisdictional challenges in proving that the alleged copyright infringement occurred within the UK.

While Getty maintained that Stability had used its images without permission, it struggled to establish a clear link between the training activities themselves, and these being carried out UK territory. Most of the training reportedly took place on servers located in the United States, outside the reach of UK copyright law. Furthermore, Getty lacked a witness who could provide a comprehensive account of the training process from start to finish, further weakening its case.

Getty also dropped its output-related copyright claims, noting that Stability had implemented technical changes to prevent the generation of images using prompts that previously produced infringing content. As a result, Getty argued that the injunctive relief it sought, had already been achieved or was no longer necessary.

What claims remain?

Despite dropping the core copyright allegations, Getty is still pursuing claims for:

• trademark infringement, particularly over the reproduction of Getty’s watermark in AI-generated images

• passing off, arguing that the use of its branding could mislead consumers

• secondary copyright infringement, based on the idea that using a model trained on infringing data could itself constitute infringement when deployed in the UK.

Implications for AI developers in the UK

This development carries important lessons for AI companies operating in or targeting the UK:

• jurisdiction matters: The case highlights the difficulty of applying UK copyright law to AI training conducted abroad

• output still counts: Even if training is lawful or outside jurisdiction, how the model is used in the UK—and what it produces—can still trigger legal risk. The remaining claims in this case could set precedent for secondary liability and trademark misuse

• policy pressure: The case has intensified calls for the UK government to clarify how copyright law applies to AI, especially in light of ongoing consultations and proposed reforms.

As the UK navigates the legal and ethical challenges of generative AI, the Getty v. Stability AI case serves as a cautionary tale, and a potential blueprint, for how future disputes may unfold.