Blow for Getty Images after UK judge removes class representative in copyright dispute with Stability AI

Case highlights challenges of bringing infringement claims against AI platforms, currently the subject of a government consultation exercise
Seattle, Washington, USA - June 26, 2016: Getty Images Sign on side of Building. Getty Images, Inc. is a visual media company, with headquarters in Seattle, Washington, United States.

Shutterstock; Eric Broder Van Dyke

Stability AI has won a key ruling in its closely watched dispute with Getty Images over alleged copyright infringement after the UK High Court removed the claimant put forward by Getty to represent the thousands of photographers who licence their content to its platform.

Legal commentators say this week’s decision by Mrs Justice Joanna Smith highlights the challenges associated with bringing copyright infringement claims against AI platforms over their use of data for training purposes.

Washington state-based photographer Thomas Barwick’s company, Thomas M Barwick, had been put forward as a class representative of the 50,000-plus photographers and content contributors whose work is exclusively licensed to Getty Images and which Stability AI is alleged to have used to train its AI model Stable Diffusion.

The other claimants comprise four Getty subsidiaries and photo library iStock, which is owned by the image licensing giant. Getty, advised by FieldFisher, had argued that it would be disproportionate to identify all members of the class owing to their “enormous number and/or the fact that the precise works used to train Stable Diffusion are within the knowledge of the defendant”.

But Stability AI, represented by Bird & Bird, maintained that Barwick did not have “the same interest” as the other copyright owners.

Siding with Stability AI, Smith said she was not persuaded that the members represented by Barwick had the same common interest as there was no definitive list of the copyright works that had been used to train Stable Diffusion and no way at present to identify the members represented by the claimant.

She added: “In the circumstances, I can see no basis on which the court can be satisfied that any particular person qualifies as a member of the class proposed or that it therefore has jurisdiction to permit this representative claim.” 

This is the latest twist in the closely watched dispute between the two parties. In December 2023, a judge in the High Court refused Stability AI’s attempts to throw out IP infringement claims brought by Getty Images.

Jennifer Heley, IP counsel at Linklaters, said: “This ongoing litigation between Getty and Stability AI illustrates broader challenges facing the English legal system in adapting traditional legal mechanisms to the evolving landscape of AI and copyright law.”

She noted that Getty had sought to avoid the “significant administrative (and potentially impossible) exercise of joining all these parties, likely to be in excess of 50,000 copyright owners, by relying on the regime in civil procedures Rules  CPR) 19.8 (representative parties)”.

She added that this decision now “limits the works Getty can rely on in this case to those it owns unless it joins relevant exclusive licensors to the case”.

She did, however, point out that Smith has left the door open for Getty to make an application under CPR 19.3 to ask the court to allow it to bring the claim without joining all relevant parties.

“Getty would need to satisfy the court that there was no risk to Stability AI of future claims (from those claimants bringing independent claims),” she explained. “Ways this could be done include Getty indemnifying Stability AI or getting undertakings from all potential claimants that they won’t sue independently.”

Gill Dennis, intellectual property law expert at Pinsent Masons, said: “It is very difficult for content creators to prove that their works have been used as training data. Even AI developers themselves do not always know with any certainty what data they have used for this purpose. This objection will, therefore, always be an obstacle to a representative claim where copyright infringement in the context of generative AI is alleged.”

In December the UK government launched a consultation on how copyrightable material can be used to train AI.

Dennis added that one of the proposals in the consultation “is to impose an obligation on AI developers to be more transparent around the training data that they use. If this proposal is accepted and implemented, it would solve the class definition issue encountered in this case”.

The case is currently pending before the High Court in London.

Email your news and story ideas to: [email protected]

Top