According to engadget, the UK High Court just handed Stability AI a significant legal victory in its ongoing copyright battle with Getty Images. Justice Joanna Smith ruled that AI models like Stable Diffusion, which don’t store or reproduce copyright works, aren’t considered “infringing copies” under UK law. This comes after Getty initially sued Stability AI back in 2023 for allegedly using millions of its protected images to train Stable Diffusion without permission. Interestingly, Getty had to withdraw its primary infringement claims because it couldn’t prove the unauthorized copying actually happened within the UK. The judge did acknowledge some evidence of Getty’s images being used—noting the presence of watermarks—but called this evidence “both historic and extremely limited in scope.” Both companies are now celebrating what they see as victories from the mixed ruling.
Sponsored content — provided for informational and promotional purposes.
What This Actually Means
Here’s the thing—this ruling doesn’t give AI companies a blank check to scrape whatever they want. The court basically said that under current UK law, the AI model itself isn’t an infringing copy because it doesn’t actually contain or reproduce the training images. But that’s a pretty narrow technical win. The judge still found evidence that Getty‘s images were used during training, which leaves the door open for other types of claims in different jurisdictions.
And let’s be real—this is exactly why Getty just struck that multi-year partnership with Perplexity AI. They’re realizing that litigation is messy and uncertain, while licensing deals provide clear revenue streams. The Perplexity agreement requires proper image attribution and linking back to source, which is basically Getty saying “if you’re going to use our stuff, at least do it the right way.”
The Bigger Picture
So where does this leave the AI copyright landscape? Basically, still messy as hell. This UK ruling contrasts with what we’re seeing elsewhere—remember the ongoing cases in the US where the outcomes might be completely different. Different countries, different laws, different interpretations.
What’s fascinating is how both sides are spinning this as a win. Stability’s general counsel says the ruling “ultimately resolves the copyright concerns,” while Getty claims it’s “deeply concerned” about the lack of transparent requirements. They’re both right in their own ways—Stability avoided the worst-case scenario, but Getty can point to this as evidence that the laws need updating.
Look, the reality is we’re watching the ground rules for AI training get written in real time through these court cases. And until there’s clearer legislation, we’re going to keep seeing this patchwork of rulings and licensing deals. The companies that can navigate this uncertainty while building ethical training practices? They’re the ones who’ll come out ahead.
