According to Financial Times News, Getty Images has lost the core of its UK copyright lawsuit against Stability AI in a High Court ruling on Tuesday that largely favored the AI company. Mrs Justice Joanna Smith ruled that Stable Diffusion doesn’t store or reproduce copyright works and therefore isn’t an “infringing copy.” Getty had abandoned its main copyright claim during the trial after finding no evidence that Stable Diffusion’s training occurred in the UK, instead pursuing narrower secondary infringement arguments. The photo agency’s shares fell about 5% following the ruling, though it did secure a win on unauthorized use of its trademarks when AI-generated images included Getty watermarks. Lawyers called the highly anticipated case a “massive damp squib” that leaves UK creators without meaningful legal precedent on whether AI training violates copyright law.
What This Means For Creators
Here’s the thing: this ruling is frustratingly narrow. It basically says that if an AI model doesn’t literally store or reproduce copyrighted material, it’s not automatically infringing. But that completely sidesteps the billion-dollar question everyone actually cares about: is training AI on copyrighted works without permission actually legal?
Iain Connor from law firm Michelmores nailed it when he called this a “massive damp squib.” Creative industries were watching this case closely because they’re genuinely worried about their livelihoods. And honestly, can you blame them? When AI companies can just vacuum up decades of creative work without compensation and then potentially compete with the very creators they trained on, that’s a pretty existential threat.
The Bigger Picture
This isn’t happening in isolation. Look at what’s been going down in the US recently – Meta won a case in June where a federal court decided using millions of books to train AI was “fair use.” Anthropic scored a similar victory. So we’re seeing a pattern here where courts are being pretty cautious about slapping restrictions on AI training.
But here’s what’s interesting: Getty did get something out of this. The court agreed that Stability AI was responsible when its systems generated images with Getty’s watermarks. That’s trademark infringement, plain and simple. And Getty’s already saying they’ll use that finding in their US case against Stability AI. So this fight is far from over.
Where Things Go From Here
The Publishers Association’s general counsel made a good point – this is “not the end of the road” for UK creators. The ruling was “hugely limited in scope” and didn’t actually rule on whether the training process itself amounts to copyright infringement. That door is still wide open.
Basically, we’re in this weird legal limbo where everyone knows these AI models were trained on massive amounts of copyrighted material, but nobody’s quite figured out the legal framework for dealing with it. Courts seem reluctant to issue broad rulings that could potentially stifle AI development, but creators are rightfully demanding some protection.
So what happens next? More lawsuits, probably. And eventually, someone’s going to have to answer the fundamental question: when does “learning” from copyrighted works cross the line into infringement? Because right now, we’ve got a whole industry built on what might be the biggest unanswered legal question of the digital age.
