Recent UK Ruling Highlights Uncertainty in How Copyright Law Applies to AI Training

As artificial intelligence tools increasingly rely on vast datasets of creative content, questions about how copyright law applies to AI training have moved to the forefront of global legal debate.

Getty Images saw most of its claims against artificial intelligence company Stability AI dismissed by a London court, which rejected the bulk of its copyright arguments and issued only a narrow ruling related to trademark use. This case is one of the first in the UK to explore how existing intellectual property law applies to generative AI and illustrates the ongoing uncertainty around how creative works are protected when used in machine learning.

Getty has alleged, through this case and others throughout the world, that Stability AI used millions of its photographs to train Stable Diffusion, a model that generates images from text prompts, amounting to copyright infringement. The company also claimed the AI’s outputs reproduced elements of its protected content.

In the course of the UK proceeding, Getty withdrew many of its allegations after being unable to provide evidence of how or where the model had been trained. The company continued to argue that Stability AI had engaged in trademark infringement and secondary copyright infringement. As a result, the court did not decide whether training an AI system on copyrighted materials violates UK copyright law. Judge Joanna Smith ruled in Getty’s favor only on instances where the Getty watermark appeared in generated images. The court dismissed the remaining copyright claims, noting that the AI system does not store or reproduce specific works.

The outcome offered Getty only very limited success, recognizing a trademark issue but leaving key copyright questions unresolved. Stability AI avoided liability on the main issue of using copyrighted materials in AI training.

As Sam Tobin reported in Reuters, this ruling points to possible gaps in UK copyright law, which predates technologies that learn from large datasets of creative content, and underscores how difficult it can be to trace and prove the use of training data.

The ruling applies only in the United Kingdom and does not affect related litigation in the United States. Most recently Getty Images (US) filed similar claims in the Northern District of California. Brought in August of 2025, in Getty Images (US) v. Stability AI  Getty has alleged:

“Stability AI has copied more than 12 million photographs from Getty Images’ collection, along with the associated captions and metadata, as well as numerous other copyrighted photographs from Getty Images’ licensees’ websites, all without permission from or compensation to Getty Images, as part of its efforts to build a competing business. As part of its unlawful scheme, Stability AI has removed or altered Getty Images’ copyright management information, provided false copyright management information, and infringed Getty Images’ famous trademarks.” 

Stability AI has moved to dismiss the complaint. That motion remains pending before the court.

The experienced trademark, copyright, and litigation attorneys at Renner Otto are happy to address any questions or concerns you may have about protecting your valuable trademark and copyright assets. Contact us to schedule a consultation.

The attorneys at Renner Otto strive to be authorities in all matters concerning the ever-evolving landscape of Intellectual Property; however, the information provided on our website is not intended to be legal advice, nor does it create an attorney-client relationship.