Getty Images sues Stability AI
Getty's UK High Court claim alleged around 11 million of its images were used to train Stable Diffusion without a licence; a separate US suit followed in February.
- Courts & copyright
- Notable
Getty Images filed proceedings against Stability AI in London’s High Court of Justice, alleging that Stability had “unlawfully copied and processed millions of images protected by copyright and the associated metadata owned or represented by Getty Images” to train Stable Diffusion, without seeking a licence Getty says it had offered. Getty’s statement said the company had approached AI developers about licensing before turning to litigation.
The UK claim, later detailed in filings, put the number of images at around 11 million and added trademark infringement and passing-off to the copyright and database-right claims: Getty argued that Stable Diffusion’s outputs sometimes reproduced a garbled version of its own watermark, which it said would mislead users into thinking Getty had produced or endorsed the image. Getty followed up on 3 February 2023 with a second suit in the US District Court for Delaware, this one alleging more than 12 million photographs copied and adding Digital Millennium Copyright Act claims over the watermark’s removal.
The two cases ran on separate tracks. The Delaware suit narrowed as it proceeded, and a US judge allowed only some claims against Stability, Midjourney, DeviantArt and Runway to continue in August 2024 in the related Andersen litigation. The UK case reached trial and, in 2025, the High Court rejected Getty’s core copyright claim — finding Stable Diffusion’s weights were not themselves infringing copies — while finding limited, historic trademark infringement over the watermark. Getty subsequently refiled elements of its case against Stability in California, continuing a dispute that, alongside Andersen, became one of the two cases through which US and UK courts worked out how copyright law applied to model training.