Contact us

Book consultation

Visit our office

Book consultation

Contact our team to find out if we can help.

Book a free call to discuss your matter with us. Please leave your details and we will call you. We would also ask you to briefly describe your matter in the notes section, for the assessment before the call.

Please kindly note, we'll try to call you within the one hour slot you book, however, sometimes we'll have to reschedule the call.

Please answer mandatory questions below.






    Contact Us

    Corporate services

    Individual services

    High Court Issues Ruling in Getty Images v Stability AI

    On 4th November 2025, the UK High Court issued its long-awaited ruling in the Getty Images vs. Stability AI case. The ruling dismisses Getty’s claim of secondary copyright infringement, while partially holding Stability AI liable for trademark infringement. This decision clarifies that training generative AI models on third-party content outside the UK is less likely to infringe UK copyright, but the use of trademarks in AI outputs remains a legal risk.

    In this article, we examine the key issues addressed by the court and clarify the implications of this ruling for AI developers, copyright holders, and the broader technology sector.

    Case Overview

    Stability AI developed Stable Diffusion, a generative AI tool that enables the creation of images from user prompts. Getty Images claimed that its images were used in training Stable Diffusion without permission, and that both the training process and generated outputs violated its copyright, database rights, and trademarks (through displayed watermarks). Getty filed suit, alleging copyright, database rights, and trademark infringement, as well as passing off.

    Getty later dropped its direct copyright and database rights infringement claims, narrowing the trial to trademark issues and secondary copyright infringement.

    Secondary Copyright Infringement

    While the majority of the court’s attention was directed at trademark infringement, the issue of secondary copyright infringement attracted significant interest.
    The key question was whether importing or downloading a pre-trained generative AI model, such as Stable Diffusion, into the UK could constitute secondary copyright infringement. The Copyright, Designs and Patents Act 1988 (CDPA) defines secondary infringement as dealing with an article known to be an infringing copy of a work. Traditionally, this applied to physical goods, but the court decided “article” could include intangible items.

    Despite this, the court rejected Getty’s argument that Stable Diffusion was an “infringing copy”. It found that an infringing copy must be a literal copy, and Stable Diffusion never stored copies of Getty’s works. Using Getty images during development in the US did not change this outcome.

    Trademark Infringement

    While the trademark infringement claims received less attention, Getty achieved partial success, albeit in a limited and historical manner. The court found evidence of double identity and the likelihood of confusion, which supported some of Getty’s claims. However, Getty’s arguments regarding trademark dilution, tarnishment, and unfair advantage were rejected. Additionally, the court did not address Getty’s claims of passing off.

    Sterling Law Comment

    When this case began, many intellectual property (IP) professionals hoped it would provide clarity on whether the current practices of generative AI developers using unlicensed third-party material to train their models were permissible in the UK. As the trial progressed, it became clear that many key questions would remain unanswered. Ultimately, the court did not address whether the use of Getty’s copyrighted works to train Stable Diffusion infringed UK copyright (since the training occurred outside the UK), nor did it resolve issues related to database rights or potential copyright infringement by the outputs.

    Getty’s secondary copyright claim failed, which likely reassures AI developers, mainly outside the UK, but disappoints UK copyright holders who worry developers can bypass UK protections by training elsewhere. This outcome adds pressure on the UK government’s ongoing consultation on copyright and AI.

    It is essential to note that this decision is final, and Getty may still choose to appeal. Additionally, this case only represents one side of the story, as Getty is pursuing a similar claim in the US, where the training and development of Stable Diffusion appears to have occurred. It will be interesting to see how that case progresses.

    See all