Getty Images v Stability AI [2025] EWHC 2863 (Ch): UK High Court rejects the secondary copyright claim and finds only 'historic' and 'extremely limited' trade-mark infringement

The problem

Generative image models such as Stable Diffusion are trained on enormous datasets scraped from the open web, and rights-holders argue that ingesting their licensed works to build a commercial model is infringement. Getty Images, which licenses stock photography, sued Stability AI in the England and Wales High Court, alleging that Stable Diffusion had been trained on millions of Getty images taken without consent and pressing copyright, database-right, trade-mark and passing-off claims (source). The case was the first major UK judgment to test how copyright and trade-mark law apply to the development of a generative-AI model, and Latham & Watkins described it as a ruling in which “the Court dismissed Getty’s secondary copyright infringement claim” (source).

What was built

Stability AI developed and released Stable Diffusion, a text-to-image generative model. As the judgment records, the model’s behaviour is fixed in its “model weights”, which are “learnable parameters controlling the functionality of the network” learned “by repeated exposure to massive quantities of data … in the form of human-generated digital images contained in datasets created by crawling and scraping images and associated descriptive captions from the Internet” (source). It was common ground at trial that “the model weights of the various Stable Diffusion versions do not store the visual information in the Copyright Works”, a fact that became central to the copyright ruling (source).

The outcome

By the time of judgment Getty had narrowed the case sharply. Mrs Justice Smith recorded that “it is now acknowledged by Getty Images that (i) there is no evidence that the training and development of Stable Diffusion took place in the United Kingdom (such that what has been called ‘the Training and Development Claim’ has been abandoned)”, and that the “Outputs Claim … has now been substantially achieved. Thus the Outputs Claim has also been abandoned” (source). Independent analysis confirmed the abandonment: Mayer Brown reported that “Getty ultimately abandoned it because Getty could not demonstrate that the acts of reproduction or storage of Getty’s images occurred within the United Kingdom during the development of Stable Diffusion” (source).

On the one copyright claim that remained, the court held that Stable Diffusion is not an “infringing copy”: the judge reasoned that “where an AI model itself is not an infringing copy, it cannot have been the intention of Parliament that it should fall within the meaning of infringing copy in section 27” and that “the secondary acts to which Getty Images objects thus fall outside the policy and object of the Act” (source). Latham & Watkins summarised the same holding: “the model weights are not themselves an infringing copy and they do not store an infringing copy” (source).

The only infringement the court found was a narrow trade-mark point: early Stable Diffusion versions had sometimes generated synthetic images bearing the Getty watermark, but the judge stressed the finding’s limits, writing that “the case is thus historic and my findings at this trial are, of necessity, extremely limited” (source). Mayer Brown’s independent analysis reported the same limit, noting that “the Court concluded that there were some ‘limited’ and ‘historic’ instances of trade mark infringement from earlier iterations of Stable Diffusion” (source). Getty, for its part, framed the result as a win on that point, stating that “today’s ruling confirms that Stable Diffusion’s inclusion of Getty Images’ trademarks in AI-generated outputs infringed those trademarks” (source).

This is a first-instance judgment under active appeal. At the December 2025 consequentials hearing “Getty Images has been granted permission to appeal the High Court’s judgment on the question of secondary copyright infringement” (source), recorded in the form-of-order ruling Getty Images (US) Inc & Ors v Stability AI Ltd (Re Form of Order) [2025] EWHC 3343 (Ch), with the judge granting “Getty permission to appeal the rejection of its secondary infringement claim” (source). As of the latest public record the appeal remained pending, with no Court of Appeal judgment issued (source). The holdings quoted above describe the un-reversed first-instance judgment and have not been overturned.

Weakest load-bearing source, named where you meet it: the courtroom facts above rest on the court’s own signed judgment (Tier 1), the primary record, not a report of it. The non-primary citations are two law-firm client alerts (Tier 2): Latham & Watkins on the secondary-copyright holding, and Mayer Brown on the abandonment and the “limited”/“historic” trade-mark finding. Both are independent expert analysis, not independent press, and each only corroborates a holding the judgment already states in full. Getty Images’ statement is the subject’s own PR characterisation (first party, non-tiered) and is quoted only as Getty’s view of the result, never as evidence of what the court decided.

Verification status: checking (amber), drafted by the war-room maker and handed to the checker. Not verified; no green badge is claimed here.

Related case files

• Thomson Reuters v Ross Intelligence: US court rejects fair use for AI training data: the US counterpart on the same question of copyrighted works as training data, decided the opposite way on the merits a US court found infringement where the UK court dismissed the surviving copyright claim on a narrower, storage-based ground.
• Bartz v Anthropic: $1.5bn AI training copyright settlement: where Getty’s copyright claim collapsed evidentially, Anthropic paid to settle a training-data copyright case, marking the commercial stakes the Getty ruling left unresolved.
• Adobe terms-of-use AI-training clarification after backlash: the contractual front of the same fight: vendors rewrite ToS on AI training precisely because rulings like Getty leave the default legal position uncertain.
• Automattic (Tumblr / WordPress) AI-training opt-out: the platform-consent analogue, where users forced an opt-out once they learned their content was being sold as training data.

Sources

1. Tier 1 (primary). High Court of Justice of England and Wales, Chancery Division, Getty Images (US) Inc & Ors v Stability AI Limited, [2025] EWHC 2863 (Ch), judgment of Mrs Justice Joanna Smith DBE, handed down 4 November 2025. National Archives caselaw: https://caselaw.nationalarchives.gov.uk/ewhc/ch/2025/2863 (retrieved live this session; local copy sources/primary-ewhc-2863.html sha1 50b986fcb97fe74c3d26686e620451beb687b639; archived https://web.archive.org/web/20260826120339/https://caselaw.nationalarchives.gov.uk/ewhc/ch/2025/2863).
2. Tier 2 (independent expert analysis). Latham & Watkins (Sophie Goossens, Brett Shandler), “Getty Images v. Stability AI: English High Court Rejects Secondary Copyright Claim,” 13 November 2025. https://www.lw.com/en/insights/getty-images-v-stability-ai-english-high-court-rejects-secondary-copyright-claim (local copy sources/latham.html sha1 2187c4ec45e27566b3d8b900ddf120c6372e8785; archived https://web.archive.org/web/20260826120639/https://www.lw.com/en/insights/getty-images-v-stability-ai-english-high-court-rejects-secondary-copyright-claim).
3. Tier 2 (independent expert analysis). Mayer Brown (Oliver Yaros, Alasdair Maher, Ellen Hepworth, Rebecca Keay, Shannon Balnaves), “Getty Images v Stability AI: What the High Court’s Decision Means for Rights-Holders and AI Developers,” 13 November 2025. https://www.mayerbrown.com/en/insights/publications/2025/11/getty-images-v-stability-ai-what-the-high-courts-decision-means-for-rights-holders-and-ai-developers (retrieved live this session; local copy sources/mayerbrown.html sha1 659129a4cc8755923c11b077d9e70cf717bcc596; archived https://web.archive.org/web/20260731011445/https://www.mayerbrown.com/en/insights/publications/2025/11/getty-images-v-stability-ai-what-the-high-courts-decision-means-for-rights-holders-and-ai-developers). Independently corroborates the abandonment of Getty’s copyright claim and the “limited”/“historic” trade-mark finding.
4. First party (subject), non-tiered. Getty Images newsroom, “Getty Images issues statement on ruling in Stability AI UK litigation,” 4 November 2025. https://newsroom.gettyimages.com/en/getty-images/getty-images-issues-statement-on-ruling-in-stability-ai-uk-litigation (local copy sources/getty-statement.html sha1 110d9e8386f18a8c27e993150549be58acc7144b; archived https://web.archive.org/web/20260826120534/https://newsroom.gettyimages.com/en/getty-images/getty-images-issues-statement-on-ruling-in-stability-ai-uk-litigation). A subject’s own statement is never Tier 1/2; quoted only for Getty’s own characterisation of the result, not as evidence of the holding.
5. Tier 2 (independent expert analysis). Burges Salmon (Harry Jewson, Emily Roberts), “Getty Images v Stability AI – Getty granted permission to appeal,” 19 December 2025. https://www.burges-salmon.com/articles/102lydc/getty-images-v-stability-ai-getty-granted-permission-to-appeal/ (retrieved live this session; local copy sources/burges-salmon-appeal.html; archived https://web.archive.org/web/20260416202420/https://www.burges-salmon.com/articles/102lydc/getty-images-v-stability-ai-getty-granted-permission-to-appeal/). Independent of counsel of record (Fieldfisher for Getty, Bird & Bird for Stability).
6. Tier 2 (independent expert analysis). The IPKat (Oliver Fairhurst), “Permission to appeal granted in Getty Images v Stability copyright claim,” 7 January 2026. https://ipkitten.blogspot.com/2026/01/permission-to-appeal-granted-in-getty.html (retrieved live this session; local copy sources/ipkat-appeal.html; archived https://web.archive.org/web/20260826163158/https://ipkitten.blogspot.com/2026/01/permission-to-appeal-granted-in-getty.html). Names the form-of-order ruling [2025] EWHC 3343 (Ch) and confirms the appeal is pending.

How this was verified

• Method: the Tier-1 judgment [2025] EWHC 2863 (Ch) was retrieved live this session from the official National Archives caselaw host, saved locally and archived to the Wayback Machine. Every quoted courtroom fact: the abandonment of the Training and Development Claim and the Outputs Claim, the “not an infringing copy” reasoning, the common-ground point that the model weights do not store the works, and the “historic … extremely limited” trade-mark finding, is quoted verbatim from that judgment. Each critical holding is now independently corroborated by a Tier-2 law-firm analysis fetched live this session: the secondary-copyright holding by Latham & Watkins, and the claim abandonment and “limited”/“historic” trade-mark finding by Mayer Brown. Getty Images’ own statement (first party, non-tiered) is quoted only as the subject’s characterisation.
• Appeal status disclosed: this is a first-instance judgment. Getty was granted permission to appeal the rejection of its secondary copyright infringement claim at the December 2025 consequentials hearing (form-of-order ruling Re Form of Order [2025] EWHC 3343 (Ch)), and as of the latest public record the appeal remained pending with no Court of Appeal judgment. Confirmed this session from two independent commentators (Burges Salmon, 19 Dec 2025; The IPKat, 7 Jan 2026), both independent of counsel of record. No holding quoted here has been reversed; the quotes describe the un-reversed first-instance judgment.
• Date verified: maker round 1, 26 August 2026; appeal disclosure added round 3, 26 August 2026, handed to the checker at status: "checking".
• What “verified” (green) would still require: the checker’s independent re-verification of every quote against the judgment, and that the appeal disclosure above remains current (no Court of Appeal reversal). No green badge is claimed here; this page is checking (amber).