
Getty Images just pulled back from some of its biggest copyright claims against Stability AI in the UK. But don’t think for a second that the war is over.
The case, filed in London’s High Court, was one of the most closely watched fights in the world of generative AI. At the center of it: Who owns the content that AI models are trained on? And can using copyrighted material to train those models be considered theft?

Getty claimed that Stability AI—creator of image generator Stable Diffusion—used millions of copyrighted images without permission to train its model. Some of the AI-generated outputs allegedly even had Getty watermarks on them.
But now, Getty has dropped two major claims:
- That Stability AI’s training process broke UK copyright law.
- That the AI’s output looked too much like Getty’s original images.
Why the retreat? According to legal experts, the UK copyright law may not cover training that happened outside the UK. Also, the outputs didn’t show enough of the original images to count as copyright violations.
Ben Maling, a lawyer at EIP, explained it simply: “Getty likely couldn’t prove the training happened in a way UK law can touch. And the generated images didn’t copy enough to count.”
Still, this doesn’t mean Stability AI is off the hook. Getty’s UK case is still alive. The company is now focusing on secondary copyright infringement and trademark violations.
The secondary claim is especially interesting. Getty says that even if the training happened abroad, using the AI model in the UK could count as “importing infringing articles.” It’s a fresh legal angle that could matter for every AI company operating across borders.
Stability AI responded by saying it was “pleased” that Getty dropped several claims. The company also believes that Getty’s remaining trademark arguments will fail, especially since users don’t see the AI-generated watermarks as messages from Stability itself.
Meanwhile, the drama continues across the Atlantic.
In the U.S., Getty is suing Stability AI separately, demanding up to $1.7 billion in damages. That case, filed in 2023, alleges the company used 12 million copyrighted images without permission. A decision on Stability AI’s motion to dismiss is still pending.
Getty has also launched its own AI image generator, trained exclusively on its licensed content. It’s a clear play to offer a “legal” alternative to tools like Stable Diffusion.
This fight isn’t just legal—it’s philosophical. In an AI world where models feed off human creativity, where’s the line between learning and stealing?
The UK retreat by Getty may be a tactical move, but the core questions remain.
And as more artists, companies, and lawmakers get involved, expect bigger battles ahead.
Also Read Microsoft May End Ties with OpenAI: What’s the Real Reason Behind the Fallout?