In the landmark case of Getty Images (US) Inc and Others v. Stability AI Ltd [2025] EWHC 38 (Ch), the English High Court addressed pivotal issues at the intersection of artificial intelligence (AI) and intellectual property law. The case centers on allegations that Stability AI, the developer of the generative AI model known as Stable Diffusion, infringed upon Getty Images’ copyrights by utilizing their protected works without authorization.
Background of the Case
Getty Images, a prominent global provider of visual content, initiated legal proceedings against Stability AI in January 2023. The crux of Getty’s claim is that Stability AI unlawfully “scraped” approximately 12 million images, along with associated metadata, from Getty’s platform. These images were allegedly used to train Stable Diffusion, an AI model capable of generating images based on textual prompts. Getty contends that this unauthorized use constitutes a direct infringement of their intellectual property rights.
Operation of Large Language Models and Copyright Implications
Large Language Models (LLMs) and generative AI systems like Stable Diffusion operate by processing vast datasets to learn patterns, styles, and structures inherent in the data. In this instance, Stability AI is accused of using Getty’s copyrighted images to train its model, enabling it to produce new images that may mimic the style or content of the original works. The central legal issue is whether such use of copyrighted material in the training phase of AI development constitutes infringement, even if the outputs are not direct replicas of the input data.
High Court’s Ruling
On January 14, 2025, Mrs. Justice Joanna Smith DBE delivered a reserved judgment addressing procedural aspects of the case. Stability AI had applied to prevent one of the claimants, Thomas M Barwick Inc, from acting as a representative for a class of individuals in the claim. The court granted Stability AI’s application, ruling that the sixth claimant could not act in a representative capacity. This decision underscores the court’s cautious approach to representative actions in complex IP litigation, particularly where the class of represented parties may not share identical interests or claims. This decision allows the case to proceed toward addressing the core issues of copyright infringement and related claims.
Parallel Proceedings in the United States
Simultaneously, Getty Images has pursued legal action against Stability AI in the United States. In August 2024, a federal judge in California ruled that visual artists could proceed with certain claims against AI companies, including Stability AI, for alleged copyright infringement. The court allowed claims related to the unauthorized storage and use of copyrighted works to train AI systems to move forward, while dismissing other claims such as unjust enrichment and breach of contract. This development indicates a judicial willingness to scrutinize the use of copyrighted material in AI training processes.
Forum Shopping Concerns
The initiation of legal proceedings by Getty Images in both the UK and the US raises questions about potential forum shopping—the practice of choosing the most favorable jurisdiction for litigation. However, considering Stability AI’s operations span multiple jurisdictions, and the alleged infringing activities occurred across borders, it is reasonable for Getty to seek remedies in both jurisdictions where the impacts are felt. Moreover, differences in copyright laws and enforcement mechanisms between the UK and the US may necessitate separate actions to fully address the alleged infringements. Therefore, while the dual proceedings could be viewed through a forum shopping lens, they also reflect a strategic legal approach to comprehensively protect intellectual property rights in a globalized digital landscape.
Merits of the Case
The core issue—whether the use of copyrighted images in AI training models constitutes infringement—remains a contentious and largely uncharted area of law. Proponents of Stability AI may argue that such use falls under fair use or fair dealing exceptions, especially if the AI-generated outputs are transformative and do not replicate the original works. Conversely, Getty Images asserts that the unauthorized use of their substantial image library for commercial AI training purposes undermines the value of their copyrighted works and violates their exclusive rights.
Given the scale of the alleged unauthorized use and the commercial implications, Getty Images presents a compelling argument that Stability AI’s actions exceed the bounds of permissible use. The outcome of this case could set a significant precedent, potentially leading to stricter guidelines and licensing requirements for the use of copyrighted materials in AI development.
It remains to bee seen
The Getty Images v. Stability AI litigation exemplifies the complex challenges at the nexus of AI innovation and intellectual property law. As courts in both the UK and the US grapple with these novel issues, their decisions are poised to shape the legal landscape for AI development and the protection of creative works going forward.
