Getty Images (US) Inc & Others v Stability AI Ltd

Quick Overview

The UK High Court ruled in favor of Stability AI Ltd in a copyright dispute brought by Getty Images, concluding that Getty failed to prove that Stability AI's training on millions of images was directly infringing or that the class action claims were suitable for representative action, thereby avoiding a potentially massive financial liability for the AI company.

Key Points: The UK High Court dismissed Getty Images' claim that Stability AI's use of millions of images to train Stable Diffusion constituted copyright infringement. The court found that Getty failed to prove that the massive scale of AI infringement claims were suitable for a representative class action under CPR Part 19.9. The judge explicitly noted that Stability AI was not a party to the contracts that granted Getty licenses to those copyrighted works. The court also rejected the representative claim because the individual claims of the 50,000+ licensees were not sufficiently similar or 'cohesive' enough to be treated as one case. Stability AI successfully argued that they did not need to provide indemnity undertakings to cover potential future lawsuits from the licensees. The judge concluded that the procedural path proposed by Getty was an 'untriable mess' and found the defendant's argument that the AI training data was not infringing to be compelling.

Context: This video discusses the January 2025 UK High Court judgment in the case of Getty Images (US) Inc & Others v Stability AI Ltd, which centered on allegations that Stability AI infringed copyright by using millions of Getty's images without consent or license to train its Stable Diffusion model. The core legal issue involved whether the sheer volume of alleged infringement could proceed as a representative class action and whether the use of copyrighted material for training constituted infringement under UK law.

Detailed Analysis

The UK High Court ruled against Getty Images in their copyright infringement claim against Stability AI regarding the training of the Stable Diffusion model on Getty's images. The court dismissed the claim for several key reasons. Primarily, the judge found that Getty failed to establish that the 50,000+ individual claims were suitable to proceed as a representative action under CPR Part 19.9, stating that the claims lacked the necessary cohesion and similarity for such a procedure. Furthermore, the judge emphasized that Stability AI was not a party to the licensing agreements between Getty and the creators, undermining Getty's ability to enforce those rights through this specific suit. The court also noted that Stability AI did not need to provide an indemnity undertaking covering future lawsuits from the licensees. Ultimately, the judge characterized Getty's proposed legal path as an 'untriable mess' and found that the defendant's argument that the use of the training data did not infringe copyright was compelling, leading to the rejection of the class action claim.