LONDON — Artificial intelligence company Stability AI largely beat Getty Images on Tuesday in the UK litigation over intellectual property.
Seattle-based Getty Images, which owns a vast online library of images and videos, sued Stability AI in a widely watched case that reached the UK High Court in June.
The case was among a wave of lawsuits filed by movie studios, authors and artists challenging tech companies' use of their works to train artificially intelligent chatbots.
Getty narrowly won its case that Stability infringed its trademark, but lost its claim of secondary copyright infringement, according to a judge's decision released Tuesday.
Both sides declared victory.
“This is a significant victory for intellectual property owners,” Getty Images said in a statement.
Getty shares fell 3% ahead of US trading
Stability said it was pleased with the decision.
“This final decision ultimately resolves the copyright issues that were the underlying issue,” said Chief Stability Counsel Christian Dowell.
Getty argued that Stability's AI image development, called Stable Diffusion, was a “brazen assault” on its image library “on a staggering scale.”
Although Getty accused Stability of copyright and trademark infringement, during the trial the company dropped its underlying allegations of copyright infringement, indicating that it did not think its arguments would succeed.
Getty also sued for trademark infringement because its watermark appeared on some images generated by its chatbot Stability.
Judge Joanna Smith said in her ruling that Getty's trademark claims are “successful (in part)” but her findings are “both historical and extremely limited in scope.”
Stability argued that the case did not apply to the United Kingdom because the training of the AI model technically took place elsewhere, on computers operated by US tech giant Amazon. He also claimed that “only a small portion” of the random output from his AI image generator is “at all similar” to Getty's work.
Tech companies have long argued that legal doctrines of “fair use” or “fair dealing” in the US and UK allow them to train their artificial intelligence systems on large volumes of text and images.
Getty is also still suing for “secondary infringement” of copyright, saying that even if the Stability AI's training took place outside the UK, offering the Stable Diffusion service to UK users would amount to importing illegal copies of its images into the country.
Smith rejected Getty's argument, saying that Stable Diffusion's AI is not infringing because it does not “store or reproduce any copyrighted works (and never has).”
Getty is also filing a copyright infringement lawsuit against Stability in the US. He originally sued Getty in 2023, but refiled the case in San Francisco federal court in August.
Getty's lawsuits are among many cases that show how the boom in generative artificial intelligence is fueling a clash between tech companies and the creative industries.
Anthropic has agreed to pay $1.5 billion to settle a class-action lawsuit by book authors who claim the company used pirated copies of their works to train its chatbot Claude.
Additionally, a federal judge dismissed the group's lawsuit. 13 authors who made similar accusations against Facebook owner Meta Platforms for training its artificial intelligence system Llama.
Warner Bros. sued Midjourney for copyright infringement, alleging that its image generator allows subscribers to create AI-generated images and videos that are protected by copyright. characters like Superman and Bugs Bunny.
Disney and Universal also sued Midjourney previously filed a separate joint copyright lawsuit alleging that the San Francisco-based startup pirated libraries to create and distribute unauthorized copies of famous characters such as Darth Vader and Minions.
___
AP Technology Writer Matt O'Brien contributed to this report.
