Stability AI Wins High Court Case Against Getty Over Copyright Infringement Allegations
In a major victory for the AI company, Stability AI has won a partial case against Getty Images in a UK High Court copyright infringement dispute. The decision comes after Getty first sued Stability AI in 2023, alleging that its Stable Diffusion AI art tool used millions of protected images from Getty without permission.
Getty had claimed that Stability AI's use of the copyrighted images for training purposes was unlawful, but the company eventually dropped its primary copyright infringement claim due to a lack of evidence. The dispute now centers on secondary infringement claims, in which Justice Joanna Smith ruled that an AI model like Stable Diffusion, which does not store or reproduce any copyrighted works, is not considered an "infringing copy" under UK law.
The High Court judge found that while Getty's images were used by Stability AI, the evidence of this was "both historic and extremely limited in scope." However, she did rule that the inclusion of Getty Images' trademarks in Stable Diffusion's outputs infringed on those trademarks. The ruling also established a precedent that intangible articles, such as AI models, are subject to copyright infringement claims.
Stability AI celebrated its win, with General Counsel Christian Dowell saying that the final ruling "ultimately resolves the copyright concerns that were the core issue." However, Getty expressed concern that even large companies can be at risk of infringement due to a lack of transparent requirements. The company urged the UK government to build on current laws around this issue.
The decision is a significant development in the debate over AI-related copyright infringement and raises questions about the responsibility of model providers for the use of copyrighted material in their training data.
In a major victory for the AI company, Stability AI has won a partial case against Getty Images in a UK High Court copyright infringement dispute. The decision comes after Getty first sued Stability AI in 2023, alleging that its Stable Diffusion AI art tool used millions of protected images from Getty without permission.
Getty had claimed that Stability AI's use of the copyrighted images for training purposes was unlawful, but the company eventually dropped its primary copyright infringement claim due to a lack of evidence. The dispute now centers on secondary infringement claims, in which Justice Joanna Smith ruled that an AI model like Stable Diffusion, which does not store or reproduce any copyrighted works, is not considered an "infringing copy" under UK law.
The High Court judge found that while Getty's images were used by Stability AI, the evidence of this was "both historic and extremely limited in scope." However, she did rule that the inclusion of Getty Images' trademarks in Stable Diffusion's outputs infringed on those trademarks. The ruling also established a precedent that intangible articles, such as AI models, are subject to copyright infringement claims.
Stability AI celebrated its win, with General Counsel Christian Dowell saying that the final ruling "ultimately resolves the copyright concerns that were the core issue." However, Getty expressed concern that even large companies can be at risk of infringement due to a lack of transparent requirements. The company urged the UK government to build on current laws around this issue.
The decision is a significant development in the debate over AI-related copyright infringement and raises questions about the responsibility of model providers for the use of copyrighted material in their training data.
! I mean, I knew this was gonna happen since Stable Diffusion is like, super popular and everyone's using it 
. But seriously, the fact that they're finally clear on not infringing on copyrighted images with their AI model is a major win for the industry 
.
. I get it, they want to protect their brands and all, but this ruling is like, super important for setting boundaries for future AI models 
. It shows that even if an AI model doesn't store or reproduce copyrighted works, it can still be affected by trademarks 
.
. And now we just gotta wait and see how the rest of the industry reacts to this new precedent 
.
and gotta say, its crazy how stability ai can just win against getty images. i mean, they did have to drop that primary claim tho... anyway, im not surprised that the court ruled in favor of stability ai though - those images are like everywhere online now. but what really got me was getty's comment about transparent requirements... yeah, thats a good point. how can we expect companies to know if their stuff is being used without permission if there aren't clear rules? and isnt this kinda the problem with a lot of tech these days? just trying to figure out who owns what when some new AI model comes along 
I don't get why they can't just get permission or something... but apparently it's a grey area now? 
So the court said no, using those pics isn't stealing if you're not actually copying them. But what about when you create something new with them, like an AI art tool? That seems kinda shady to me 
.
does this mean that companies are gonna start using more transparent requirements for their training data? and what about the people behind these models - are they responsible for making sure their trainings don't infringe on copyrights?
... anyway, back to AI and stuff. So, this whole copyright thing is making me think about how we're basically relying on these massive databases of images to train our AI models, which is kinda wild 
.
. As someone who's been around for a bit, I think it's interesting that the judge ruled that intangible articles like AI models can be subject to copyright infringement claims 
, but what really concerns me is how Getty is saying it could affect big companies too. I mean we don't want some huge corporation getting taken down over something as complex as AI law 
. It'll be interesting to see how this all plays out in the future 
. And yeah, the part about trademarks being infringed on is a bummer, but it's not like Stability AI was intentionally trying to mess with Getty's brand 
. It's all just about navigating these complex laws and figuring out what's fair and reasonable 
. Now we can start talking about the bigger picture - how do we regulate this stuff? Should models be more transparent about their training data? 
But also, can we make it easier for people to figure out what's good and bad when it comes to using copyrighted images in your training data? That feels like a pretty important conversation to have.
plus, the judge said that stability ai doesn't store or reproduce copyrighted works, which is pretty standard for AI models nowadays 
. But at the same time, this case also raises questions about responsibility and liability for AI model providers. Are they really liable for what their models do? It's a mess, but I guess that's just how it goes sometimes 
.
.
. It's like, if you're using someone else's library book as research material, is it really that different from using their images in an AI model? 
. Anyway, stability wins this round 
, but I'm sure there are more battles to be fought in the world of AI copyright law 