Artists file a class action over image generators

Three illustrators — Sarah Andersen, Kelly McKernan and Karla Ortiz — filed a putative class action against Stability AI, Midjourney and DeviantArt in the US District Court for the Northern District of California. It was the first generative-image lawsuit brought by working artists themselves, rather than by a rights-holding company such as Getty Images, which sued Stability days later.

The complaint argued that Stable Diffusion, the Midjourney product, DreamStudio and DeviantArt’s DreamUp tool were all trained on billions of images scraped from the web — including the plaintiffs’ copyrighted work — without licence, credit or compensation, and that the resulting models were “21st-century collage tools” capable of producing images in a named artist’s style on demand. Alongside direct and secondary copyright infringement, the suit raised claims under the right of publicity and California’s unfair competition law, and sought to represent a class of artists whose work appeared in the LAION training dataset.

The case attracted attention less for its legal novelty — training-data disputes were already emerging elsewhere — than for putting named creators, rather than a corporate plaintiff, at the centre of the argument that generative image models compete with the artists whose work trained them.

A California federal judge dismissed most of the original claims in October 2023 with leave to amend; the plaintiffs added further artists and Runway AI as a defendant. In August 2024, Judge William Orrick allowed the core copyright claims against Stability, Midjourney, DeviantArt and Runway to proceed, making Andersen v. Stability AI one of the cases setting the terms on which US courts would treat AI training as fair use — or not.