IP Law Daily, TRADEMARK—N.D. Cal.: Stock photo library states infringement claims over watermarks in AI-generated images, (Apr 24, 2026)

By Mina O. Capouet, J.D., LL.M.

In the court’s view, Getty Images sufficiently alleged trademark infringement and unfair competition over use of images to train generative AI models, but not DMCA claim.

Stock imagery company Getty Images sufficiently alleged claims for trademark infringement and unfair competition against Stability AI over related to the use of Getty’s copyrighted imagines to train AI image-generation models, the U.S. District court for the Northern District of California has decided. Getty’s distorted watermarks, which appeared on images generated by Stability’s AI tool could mislead consumers into thinking that Getty was affiliated with or endorsed those images. The AI-generated images with misleading watermarks could confuse consumers as to origin or endorsement. However, Getty failed to assert a violation of the Digital Millennium Copyright Act. The court denied Stability AI’s motion to dismiss Getty’s trademark and unfair competition claims, but it granted the motion with respect to the DMCA claim (Getty Images (US), Inc. v. Stability AI, Ltd., No. 3:25-cv-06891-TLT (N.D. Cal. Apr. 23, 2026)).

Getty Images, Inc., a stock photography company, curates a collection of premium quality visual assets and licenses them to a commercial customer base consisting of media outlets, advertising agencies, and corporations. Stability AI, Ltd. (“Stability AI”) is the creator of image-generating models, including “Stable Diffusion,” which uses artificial intelligence to generate computer-synthesized images in response to user-supplied text prompts. Getty alleged that Stability AI’s models were trained on data from a non-party that created an open dataset of billions of hyperlink-text pairs scraped from the public internet. Getty further claimed that the images generated by its models contain a distorted version of its watermark, thus creating confusion as to the source of the images and falsely implying an association with Getty.

Section 1202(a) of the DMCA. Getty alleged that Stability AI violated Section 1202(a) of the DMCA by including false copyright management information (CMI) in its AI-produced images—that is, the distorted watermarks. In the court’s view, the complaint lacked allegations to suggest that Stability AI acted with specific intent to induce, enable, facilitate, or conceal copyright infringement. Therefore, the court found that Getty failed to state a claim under Section 1202(a). To pass muster, the party must adequately allege a double scienter, (1) that the infringement was knowing and intentional and (2) there was specific intent to induce, enable, facilitate, or conceal the infringement. In this instance, Getty adequately pleaded that Stability AI knowingly provided false CMI of Getty, however, it did not adequately allege anything with respect to Stability AI’s intent to induce or conceal infringement through use of its copyrighted information.

Trademark infringement claims. The court denied Stability AI’s motion to dismiss Getty’s claims that use of its mark is likely to cause confusion under 15 U.S.C. § 1114(1). The court found that Stability AI’s displays of Getty’s marks on outputs of their models, which are made available to potential purchasers of competing generative AI services was adequate to allege that the marks are used in commerce and that it would create a likelihood of consumer confusion. Because Getty had established widespread recognition and the images generated by Stability AI’s models directly compete with Getty’s images, including instances in which users of Stable Diffusion reported that image output created a distorted version of Getty’s mark, the court found that it would be reasonable to assume that consumers would be confused to some extent by the similar imagery.

False designation of origin claims. Finding that Getty’s claim for false designation of origin in violation of 15 U.S.C. § 1125(a) (Lanham Act Section 43(a)) was not barred by Dastar Corp. v. Twentieth Century Fox Film Corp., 539 U.S. 23 (2003), the court denied the motion to dismiss this claim. The court rejected Stability AI’s argument that Getty’s watermark was “part of the communicative content that Dastar” held that the Lanham Act does not protect. Finding that Stability AI’s images indicate the “origin of goods” under the Lanham Act, the court stated that Dastar does not bar Getty’s false designation of origin claim as the claim in question was a “passing off” as Stability AI distributed images generated by their models containing distorted Getty watermarks without attribution to Getty. Additionally, the court noted that Getty had alleged that the distorted watermark would confuse consumers regarding Getty’s relationship with Stability AI, thereby diminishing the mark’s value as well as several instances of Stability AI’s unauthorized use of Getty’s expressive content.

Trademark dilution claims. The court denied Stability AI’s motion to dismiss Getty’s claims under 15 U.S.C. § 1125(c)(2)(A), as all facts weighed together, Getty had alleged a “facial plausibility that the general public actually recognizes its works.” The court found that Getty had met the burden of pleading that its mark held “household” status by alleging that it garnered ‘“2.8 billion searches annually” on its websites in “23 languages”’ and appeared every day in the world’s most influential newspapers, magazines, films, and books, among other things.

California unfair competition law claim. The court denied Stability AI’s motion to dismiss Getty’s claims under California’s unfair competition law, which broadly prohibits any unlawful, unfair, or fraudulent business practices under a three-pronged statute. The court explained that the “unfair prong” is met by showing that the conduct is comparable to or the same as a violation of the law or significantly threatens competition and that a plaintiff must demonstrate actual reliance as to the “fraudulent prong.” The court found that Getty sufficiently alleged its trademark claim under 15 U.S.C. 1114(1) whereby Stability AI’s conduct resulted in an unfair advantage by misappropriating its reputation to compete and that the complaint went beyond the classic trademark or copyright law claim by alleging misrepresentation, whereby it was likely Getty’s clients would mistakenly believe there was an affiliation with Stability AI and Getty. The court found that given these allegations, the three prongs had been met.

Conclusion. The court denied Stability AI’s motion to dismiss Getty’s trademark and unfair competition claims. The court dismissed the DMCA claim without prejudice and granted leave to amend.

The Case is No. 3:25-cv-06891-TLT.

Judge: Thompson, T.

Attorneys: Benjamin E. Marks (Weil, Gotshal & Manges LLP) for Getty Images (US), Inc. Joseph Charles Gratz (Morrison & Foerster LLP) for Stability AI, Ltd.

Companies: Getty Images (US), Inc.; Stability AI, Ltd.

Cases: AINews Copyright Trademark TechnologyInternet CaliforniaNews GCNNews