In Getty Images (UK) and others v Stability AI [2025] EWHC 2863 (Ch), photo agency Getty sought to protect its millions of high-quality photographic images and alleged Stability scraped those images to train its AI model, Stable Diffusion, without consent. However, the claim faced jurisdictional issues as Getty could not prove the training took place in the UK. Getty also scaled back its claim as Stability had blocked prompts used to generate images that would lead to primary infringement.
Luke Maunder, partner at Osborne Clarke, said the decision did not address the ‘core issue of the alleged primary copyright infringement by the training of AI models.
‘The field is open and we may still see government policy or legislation before a case tries to cut that Gordian Knot’.
Ellen Keenan-O'Malley, solicitor at EIP, said: ‘From a copyright law perspective, this case ended up being a damp squib.’
Handing down judgment this week, Mrs Justice Joanna Smith held Stability breached Getty’s trade mark by reproducing its watermark on generated images but dismissed Getty’s secondary infringement claim.
James Clark, partner at Spencer West, said: ‘At the end of the training process, the AI model did not store any copy of the protected works, and the model itself was not itself an infringing copy of such work.
‘It is this finding that will cause concern for the creative industry while giving encouragement to AI developers.
‘The judgment usefully highlights the problem that the creative industry has in bringing a successful copyright infringement claim in relation to the training of large language models. During the training process, the model is not making a copy of the work used to train it, and it does not reproduce that work when prompted for an output by its user.’
Nathan Smith, IP partner at Katten Muchin Rosenman, said: ‘On the face of it, the judgment appears to present a win for the AI community, but arguably leaves the legal waters of copyright and AI training as murky as before.’
