Stability AI Can’t Escape Getty Images Copyright Lawsuit in UK, Court Says

Image: Stability AI

Stability AI Can’t Escape Getty Images Copyright Lawsuit in UK, Court Says

A copyright battle over AI-generated imagery will go to trial in the United Kingdom, a court in London has ruled. Getty Images filed a lawsuit against Stability AI early this year for allegedly making unauthorized use of Getty-owned images to train its AI-powered models. In ...

December 5, 2023 - By TFL, Miriam Gemmell

Stability AI Can’t Escape Getty Images Copyright Lawsuit in UK, Court Says

Image : Stability AI

Case Documentation

Stability AI Can’t Escape Getty Images Copyright Lawsuit in UK, Court Says

A copyright battle over AI-generated imagery will go to trial in the United Kingdom, a court in London has ruled. Getty Images filed a lawsuit against Stability AI early this year for allegedly making unauthorized use of Getty-owned images to train its AI-powered models. In response, London-based Stability AI sought to escape the lawsuit, arguing that the UK is not the proper forum, as the large language models that underpin its Stable Diffusion text-to-image generator were not trained – or developed – in the UK. Refusing to toss out the case, a High Court justice held last week that the case will proceed to trial to determine the extent to which any development and training took place in the UK.  

Taking issue with some of the evidence provided by Stability AI, Mrs. Justice Joanna Smith of the England and Wales High Court held in her December 1 order that evidence provided by Stability AI founder and chief executive officer Emad Mostaque regarding where the relevant AI models were developed and/or trained may be “inaccurate, or incomplete,” with the judge noting that “apparent inconsistencies between Mr. Mostaque’s evidence and other available evidence casts doubt over the credibility of his statements generally.” Among other things, Mostaque previously told the court that he is “confident that no Stability AI employee based in the UK has ever worked on developing or training Stable Diffusion.” 

(Note: Counsel for Stability AI also argued in furtherance of its motion strike claims and/or grant summary judgment against Getty Images that Stable Diffusion does not contain the images said to have been used to train it, and in any event the model is not an ‘article’ within the meaning of the relevant legislation (s.27 of the Copyright, Designs and Patents Act 1988), and that the appearance of the Getty Images watermark on some of the images  (as the associated trademarks) does not amount to “use in the course of trade” of those marks and/or is the result of the prompt used not the model itself.)

Mrs. Justice Smith contrasted Mostaque’s statements with those that he made in previous interviews, such as a comment in which he revealed that Stability AI “had ‘fast tracked’ the residency applications of Ukrainian and Russian developers to allow them to come to the UK,” as reported by the Telegraph. This statement and others could mean that “developers [were] working in the United Kingdom in connection with the development and training of Stable Diffusion,” which would have a bearing on the appropriateness of trying the case in the UK. 

In light of such “unanswered questions,” the judge held that the issue of where the underlying models were developed and/or trained could not be decided at the summary judgement stage and must be decided at trial. 

A Bi-National Battle

Not the only case that pits Stability AI against Seattle-headquartered Getty Images, as the latter filed a headline-making lawsuit against the generative AI platform provider in a federal court in Delaware in February. In its stateside suit, Getty claims that as part of a “brazen infringement of [its] intellectual property on a staggering scale,” Stability AI has copied millions of photographs from its collection “without permission from or compensation to Getty Images, as part of its efforts to build a competing business.” 

In addition to setting out copyright infringement claims in connection with Stability AI’s alleged “scraping [of] links to billions [images] from various websites, including Getty Images’ websites,” which are “highly desirable” for training AI programs, Getty also makes trademark infringement and dilution claims. According to Getty, “the Stable Diffusion model frequently generates output bearing a modified version of the Getty Images watermark,” which appear, in some cases, on output that is “of much lower quality and at times ranges from the bizarre to the grotesque.” 

Beyond causing confusion (and thereby, giving rise to trademark infringement), Getty asserts that Stability AI’s “incorporation of Getty Images’ marks into low quality, unappealing, or offensive images dilutes those marks in further violation of federal and state trademark laws.”

Stability AI has sought dismissal in that case, arguing this spring that the court lacks personal jurisdiction over its UK-based arm Stability AI Ltd., which Getty added as a defendant when it amended its complaint in March. Part of the problem, according to Stability AI, is that the photo agency has “not even attempted to plead a statutory basis for jurisdiction” for Stability AI Ltd., which is a “necessary and indispensable party.” In short: “The Amended Complaint fails to allege any connection between Delaware and Stability AI Ltd.’s alleged infringing acts,” and thus, the court has “two options: dismissal or transfer.” 

As for Getty’s likelihood of success in the UK versus the U.S., it may have an advantage on Stability AI’s home turf, as “the UK’s ‘fair dealing’ exceptions to copyright infringement are narrower than the more fluid ‘fair use’ exceptions in the U.S.,” Potter Clarkson’s Mark Nichols stated in a note. While the UK’s exception to copyright infringement for text and data mining is currently restricted to certain uses, such as for non-commercial research and/or “criticism, review or quotation,” Nichols asserts that “the UK government has proposed to extend this exception to text and data mining of copyright materials … which could enable generative AI to be trained on online data without risking copyright infringement.” 

The UK government has since revealed that it will not proceed with “a broad copyright exception for data mining,” but a representative for the UK’s Intellectual Property Office said that it “seeks to strike a balanced and pragmatic approach through the development of a code of practice which allows AI innovators and the creative industries to grow in partnership.” This includes “making licenses for data mining more available,” helping “to overcome barriers that AI firms and users currently face, and ensuring that there are protections for rights holders.” 

The case is Getty Images (US), Inc. v. Stability AI, Inc., 1:23-cv-00135 (D.Del.).

related articles