California Court Lets Getty’s Trademark Claims vs. Stability AI Move Forward
A federal judge ruled Getty Images’ trademark and unfair competition claims against Stability AI can proceed.
Why it matters: Legal professionals dealing with AI and intellectual property should note that the court signaled trademark law can apply to AI training practices. The decision sets an early precedent on brand protection where AI uses trademarked content without authorization.
- On April 23, 2026, the Northern District of California let Getty's trademark, dilution, and unfair competition claims proceed.
- Getty says Stability AI generated images featuring distorted versions of its watermark, potentially confusing consumers.
- The court found Getty’s marks are famous, citing 2.8 billion yearly searches across 200 countries.
- Getty’s DMCA claim was dismissed for lack of specific intent to induce or conceal infringement.
The U.S. District Court for the Northern District of California ruled April 23, 2026, that Getty Images’ trademark and unfair competition claims against Stability AI are sufficiently pled to survive dismissal. Getty accuses Stability AI of using its images in training and producing AI-generated content containing warped Getty watermarks—raising the risk of consumer confusion about origin.
- Getty’s claims under trademark infringement, false designation of origin, dilution, and California’s Unfair Competition Law will move forward. The court found Getty’s allegations “sufficient to pass muster at this stage.”
- The opinion highlights that Getty’s marks are “famous”, pointing to metrics like over 2.8 billion annual website searches in 23 languages and a global client base of more than 708,000 across 200 countries.
- The court also allowed claims of unlawful, unfair, and fraudulent business practices under state law, based on Getty’s allegations about Stability AI’s model outputs and use of trademarks.
- Getty’s DMCA claim was dismissed, with the court finding Getty failed to allege the required specific intent for false copyright management information: “The court rejected Getty’s contention that the requisite intent can be 'plausibly inferred' at the pleading stage.”
This case signals that IP owners may be able to assert trademark rights against AI companies using content for model training—expanding the legal landscape for both rights holders and developers.
By the numbers:
- 2.8 billion — annual searches on Getty Images’ website in 23 languages
- 708,000 — number of Getty Images customers across 200+ countries
Yes, but: The court rejected Getty’s DMCA claim, highlighting the need for specific intent that could not be simply inferred at this stage.