Getty Images largely loses suit over AI image generator
Getty Images has largely lost a landmark case against AI company Stability AI in the UK over its image generator.
Seattle-based Getty, which produces editorial content and creative stock images and videos, accused Stability AI of using its images to “train” its Stable Diffusion system, which can generate images from text inputs.
The company had sued Stability AI for copyright infringement, on the grounds that Stable Diffusion was trained using images from Getty and that the images created by Stable Diffusion reproduced copyrighted images.
However, Getty dropped this part of his case midway through the hearing, partly due to a lack of evidence as to where Stable Diffusion was “trained”; This could limit the broader significance of Tuesday’s decision on artificial intelligence legislation, intellectual property lawyers said.
Getty’s trademark infringement and secondary copyright infringement claims remained in play before the London High Court’s ruling, alleging that Stability AI had imported a copyright-infringing AI model into the UK.
In her written decision, Judge Joanna Smith said Getty was “partially” successful on the trademark infringement issue, but her findings were “extremely limited in both history and scope.”
It also rejected Getty’s claim of secondary copyright infringement.
Getty Images said in a statement that the decision “confirms that Stable Diffusion’s inclusion of Getty Images’ trademarks in AI-generated output infringes those trademarks.”
The statement said the decision “establishes a strong precedent that intangible articles, such as AI models, are subject to copyright infringement claims in the same way as tangible articles,” and Getty said it would use it in a parallel lawsuit against Stability AI in the United States.
Gill Dennis, a solicitor at Pinsent Masons, said the decision meant the government must now provide “clear and timely policy guidance” on the question of whether training AI on copyrighted material is an act of infringement.
Lawyers said Getty’s decision to abandon most of its case meant that the issue of whether training AI models on copyrighted works constituted infringement in the UK remained unresolved.
“It leaves the UK without a meaningful decision on the legality of the process by which an AI model learns from copyright material,” said Iain Connor of law firm Michelmores.
