Getty Images has filed a case against Stability AI, alleging that the company copied 12 million images to train its AI model ‘without permission … or compensation.’
Getty Images has filed a lawsuit in the US against Stability AI, creators of open-source AI art generator Stable Diffusion, escalating its legal battle against the firm.
The stock photography company is accusing Stability AI of “brazen infringement of Getty Images’ intellectual property on a staggering scale.” It claims that Stability AI copied more than 12 million images from its database “without permission … or compensation … as part of its efforts to build a competing business,” and that the startup has infringed on both the company’s copyright and trademark protections.
The lawsuit is the latest volley in the ongoing legal struggle between the creators of AI art generators and rights-holders. AI art tools require illustrations, artwork, and photographs to use as training data, and often scrape it from the web without the creator’s consent.
Getty announced last month that it has “commenced legal proceedings in the High Court of Justice in London” against Stability AI. However, that claim has not yet been served, and the company did not say at the time whether or not it also intended to pursue legal action in the US. Stability AI is also being sued in US along with another AI art startup, Midjourney, by a trio of artists who are seeking a class action lawsuit.
“We can confirm on Friday Getty Images filed a complaint against Stability AI, Inc. in the United States District Court in Delaware,” Anne Flanagan, vice president of communications at Getty Images, told The Verge. “Getty Images has also filed a Claim in the High Court, which has not been served at this time. As is customary in the UK, on January 16 Getty Images sent and requested a response to a letter before action from Stability AI Limited within a customary timeframe. Stability AI Limited have confirmed receipt of this letter.”
Legal experts say Getty Images’ case is on stronger footing than the artist-led lawsuit, but caution that in such unknown legal territory it’s impossible to predict any outcome.
Andres Guadamaz, a UK academic specializing in AI and copyright law, said Getty’s complaint was “very strong,” on Twitter. “The complaint is technically more accurate than the class action lawsuit,” said Guadamaz. “The case will likely rest on the [copyright] infringement claim, and the defendants are likely to argue fair use. Could go either way.”
Aaron Moss, a copyright lawyer at Greenberg Glusker and publisher of the Copyright Lately blog, tweeted: “Getty’s new complaint is much better than the overreaching class action lawsuit I wrote about last month. The focus is where it should be: the input stage ingestion of copyrighted images to train the data. This will be a fascinating fair use battle.”
Moss was the first to publish the full complaint on his blog. The suit has not yet been given a case number and at the time of publication does not yet appear on the official online index of federal court documents, PACER. However, a representative for Getty Images confirmed to The Verge that the document shared by Moss is accurate.
Speaking to The Verge via DM, Moss noted that the would-be class action lawsuit “was much more focused on the occupational harm caused to working artists by the proliferation of AI tools,” while Getty’s concentrates “on the fact it wasn’t paid for the use of its images.” Notably, Getty has licensed its images and metadata to other AI art generators, underscoring the fact that Stability AI willfully scraped its images without permission.
The copyright infringement arguments in the lawsuit will turn on the interpretation of the US fair use doctrine, which protects the unlicensed use of copyrighted-work in certain scenarios. The concept of “transformative use” is also likely to be an important factor. Is the output of Stable Diffusion different enough from its training data? Recent research has found that the software memorizes some of its training images and can reproduce them almost exactly, though this only happens in a very small number of cases.
Another argument floated by Getty Images relates to its trademark. Stable Diffusion is well known for recreating the company’s watermark in some of its images, and Getty argues that the appearance of this watermark on the model’s often “bizarre or grotesque images, dilutes the quality of the Getty Images Marks by blurring or tarnishment.”
The case will be slow to move forward though, cautioned Moss. He notes that it was filed in the District Court of Delaware, and that the court’s docket is “pretty backed up.”
“I’m currently handling a matter there, and was told that judges routinely take months (like sometimes up to 6-9 months) to decide motions to dismiss after they’re submitted,” Moss told The Verge. “it will likely take several years for the Getty Images case to get through discovery and summary judgment motions before trial.”
He notes that such fair use cases also require input from both judges and juries. “The jury decides any disputed factual issues, but the ultimate legal questions are supposed to be decided by a judge,” says Moss.
The Verge has reached out to Stability AI for comment and will update this story if we hear back.
You can read Getty Images’ complaint (23-cv-135) in full below: