Getty v. Stability (US)

The claims

Copyright infringement, DMCA false copyright-management-information, trademark infringement, false designation of origin, trademark dilution, unfair competition (Cal. UCL)

What has happened

Getty filed the US counterpart to its London case in Delaware in February 2023, alleging Stability copied more than 12 million Getty photographs, along with captions and metadata, to train Stable Diffusion, and that outputs reproduced distorted Getty watermarks. The Delaware action spent over two years in jurisdiction and venue fights. On August 14, 2025, Getty voluntarily dismissed it and refiled the same day in the Northern District of California, adding allegations that Stability’s conduct dilutes the market for Getty’s licensed images. In April 2026 the court denied Stability’s motion to dismiss the trademark infringement, false designation, dilution and unfair competition claims, dismissing only the DMCA false-CMI claim, without prejudice, for failure to plead intent. Stability did not challenge the copyright claim at that stage, and the case moved into case management and discovery.

Key developments

• 2023-02-03 — Getty sued Stability AI in the District of Delaware over the training of Stable Diffusion on its image library.
• 2025-08-14 — Getty voluntarily dismissed the Delaware action and refiled in the Northern District of California, adding market-dilution allegations.
• 2026-04-23 — The court denied Stability’s motion to dismiss the trademark, false designation, dilution and unfair competition claims; the DMCA false-CMI claim was dismissed without prejudice for inadequately pled intent.
• 2026-05-07 — Case management proceedings continued in San Francisco, with a further conference set for November 2026 and no trial date reported.

Why it matters for training data

This is the US case that squarely pairs a large licensor’s copyright claims with trademark claims over watermark artifacts in outputs. The April 2026 ruling means an AI developer can face Lanham Act and state unfair-competition exposure for output artifacts even before any fair-use question is decided. For data suppliers, it supports the position that scraped stock libraries carry both copyright and brand risk; for model buyers, watermark contamination in outputs is a concrete diligence item.

Sources

• Bloomberg Law on the Delaware dismissal and refiling
• IP/DE on the voluntary dismissal
• Loeb & Loeb summary of the April 2026 ruling
• CourtListener docket (N.D. Cal.)

Deeper analysis

• Training-data copyright lawsuits: the landscape
• Is scraped data fair use for AI training?

← All case files Weekly lawsuit news