Hot off the Press – a UK High Court decision has now clarified when content owners can bring copyright claims against AI tool developers – where those tools are trained on publicly available content.
The trial of Getty’s case – which is one of several lawsuits brought in the UK, the US and elsewhere over the use of copyright-protected material to train AI models and that we have previously reported on here and here – began in June this year with much anticipated about its potential wider impact on the law on AI.
In the case, Getty argued that Stability AI used millions of its images without permission to train their AI system “Stability Diffusion”. Getty said this violated its copyright in those images and also harmed its brand – because some AI-generated images contained distorted versions of the Getty watermark.
However, the court dismissed Getty’s main copyright claims. The reason was that Getty could not show that the alleged copying of its images happened in the UK, which is required for primary copyright infringement under UK law. Stability AI did admit that some Getty images were included in training data, but the precise location of the copying was never proven.
Getty then focused on secondary copyright infringement, which concerns activities like importing, selling, or distributing material that is already considered an unlawful copy. For these claims to succeed, Getty had to show that Stable Diffusion itself comprised an “infringing copy” of Getty’s works. The judge accepted that software in digital form can count as an “article” under the law but ruled that Stable Diffusion does not amount to a copy of Getty’s images on the grounds that “Stable Diffusion… does not store or reproduce any copyright works”. Therefore, the secondary infringement claims were also rejected.
However, Getty did succeed in part of its trade mark case. Tests showed that some images produced by Stable Diffusion still displayed the Getty watermark, even though the images were not real Getty photos. The court agreed that this could mislead users and damage Getty’s brand, and therefore this aspect of the claim was upheld.
Getty said the ruling highlights how difficult it is for creators to defend their work when AI companies do not have to reveal how training data is collected. As a result, Getty has called for stronger transparency rules and clearer legal protections. Stability AI also welcomed the ruling and said it resolves major copyright concerns raised in the case.
This decision comes as the UK Government considers how to update copyright rules to respond to AI technology. Creative industries are pushing for tighter protections and fair payment when their work is used to train AI. Meanwhile, AI technology companies are arguing for more freedom to use data, saying that restricting AI development could hold back innovation and economic growth. The government is expected to publish more detailed proposals later this year.
Scintilla is Where Innovation Thrives
We put you in control of your destiny by organising, optimising, & monetising your intellectual property.