Getty Images (US), Inc. v. Stability AI, Ltd.

In Getty Images’ lawsuit against generative AI company Stability AI, district court finds that Getty adequately pled claims for trademark infringement, false designation of origin and trademark dilution under federal law and unfair competition under California law but that plaintiff did not adequately plead scienter requirement for its false copyright management information claim under DMCA.

Getty Images (US) Inc., a stock photography company, sued Stability AI Ltd., Stability AI Inc. and Stability AI US Services Corp., asserting several different causes of action under federal and California state laws arising out of Stability AI’s alleged infringement of Getty’s intellectual property, including copyright infringement, false copyright management information (CMI) under the Digital Millennium Copyright Act (DMCA), trademark infringement, false designation of origin, trademark dilution under federal law and unfair competition under California law. Getty alleged that Stability AI’s image-generating models use artificial intelligence (AI) to generate computer-synthesized images in response to user prompts and that those images contain distorted versions of Getty’s watermark, which creates confusion as to the source of the images and falsely implies Getty’s association with them. Leaving aside Getty’s copyright infringement claim, Stability AI moved to dismiss the rest of the claims.

The district court denied defendants’ motion as to all claims except the false CMI claim. Under the DMCA, a plaintiff must plead actual knowledge that the CMI at issue is false and that the defendant intended to induce, enable, facilitate or conceal infringement. While the court found that Getty adequately pled that Stability AI knowingly provided Getty’s false CMI, Getty did not plead that Stability AI acted with the requisite intent to further or conceal infringement. The court rejected Getty’s contention that the requisite intent can be “plausibly inferred” at the pleading stage. On that basis, the court dismissed Getty’s false CMI claim without prejudice.

As to Getty’s trademark infringement claim, Stability AI argued that Getty failed to allege both “use in commerce” and a plausible likelihood of confusion. The court concluded that Getty adequately pled both elements, because it alleged that its mark has been in use in the U.S. since 1995 and that images generated by Stability AI’s models compete with Getty’s images. Getty also cited sources showing the user counts for these models, the number of images generated by these models and the market’s transition to AI-generated images. It further alleged that users of Stability AI’s models reported that their image output created a distorted version of Getty’s trademark. The court found these allegations “sufficient to pass muster at this stage.”

As to Getty’s false designation of origin claim, Stability AI argued that it is a repackaged copyright infringement claim in trademark clothing, which is barred under the U.S. Supreme Court’s 2003 decision in Dastar Corp. v. Twentieth Century Fox Film Corp. The court disagreed, explaining that Getty’s claim is not barred by Dastar because Stability AI’s images are “origin of goods” under the Lanham Act, and Getty alleged several incidents showing that the distorted Getty watermark on the images produced by Stability AI’s models would confuse customers about Getty’s relationship with Stability AI.

Moving on to Getty’s trademark dilution claims, the court noted the requirement that a plaintiff plead facts showing that its marks are “famous” and, after delineating four factors that a court may consider, found that Getty had done so. In particular, Getty alleged that its name and trademarks are renowned in the U.S. and around the world, with 2.8 billion searches on its website annually in 23 languages, and that it served more than 708,000 customers in more than 200 countries last year. Getty’s images appear every day in the world’s most influential newspapers, magazines, advertising campaigns, films, television programs and books and on websites. The court noted that Getty owns eight trademarks on the Principal Register of the U.S. Patent and Trademark Office and explained that, even though Getty did not allege the extent to which consumers recognize its mark, when the court weighed the facts alleged together, it found that Getty created at least a “facial plausibility that the general public actually recognizes its mark.”

On Getty’s unfair competition claim, the court explained that a California Unfair Competition Law (UCL) claim borrows violations of other laws and treats them as unlawful practices that the UCL makes independently actionable and, because Getty adequately alleged its trademark infringement claim, it adequately pled the unlawful prong under the UCL. The court also found that Getty pled the unfair and fraudulent prongs by alleging that Stability AI’s conduct gives it an unfair advantage by using Getty’s marks in “competitive proximity” and misappropriating its reputation and that a substantial portion of Getty’s current and potential clients are likely to falsely believe that Stability AI is affiliated with Getty.

Summary prepared by Tal Dickstein and Jennifer Kahn