AI platforms score early win in copyright class action

Managing IP is part of Legal Benchmarking Limited, 4 Bouverie Street, London, EC4Y 8AX

Copyright © Legal Benchmarking Limited and its affiliated companies 2025

Accessibility | Terms of Use | Privacy Policy | Modern Slavery Statement

AI platforms score early win in copyright class action

Stability AI.jpeg

A California court dismissed most of the claims filed by a group of artists in a copyright case against Midjourney, Stability AI and DeviantArt

The US District Court for the Northern District of California dismissed most of the claims filed by a group of artists against three generative artificial intelligence platforms in a copyright case yesterday, October 30.

Artists Sarah Anderson, Kelly McKernan, and Karla Ortiz accused Midjourney, DeviantArt, and Stability AI of infringing their artwork.

In a class-action complaint filed in January this year, the artists alleged that the platforms had taken billions of training images that had been scraped from public websites, including their own. The platforms subsequently filed a motion to dismiss the claims.

Judge William Orrick dismissed the complaint against DeviantArt and Midjourney ruling that it “was defective in numerous aspects”.

Orrick allowed Anderson to pursue a claim that Stability AI had used her works for training.

However, he dismissed McKernan and Ortiz’s claims against Stability AI because they had not registered their work with the US Copyright Office – a prerequisite for bringing an infringement action in the US.

The judge also dismissed right of publicity and unfair competition claims brought against the three platforms.

Advantage AI

Earlier this year, when various copyright infringement lawsuits were filed against generative AI platforms, counsel predicted that the biggest challenge for copyright owners would be fending off the platforms’ motions to dismiss.

In his reasoning for siding with the AI platforms, Orrick noted that the plaintiffs admitted it was likely that the platforms' images would be unlikely to match with any specific image they had created.

The artists were, however, granted leave to amend their claims, which means they can rethink their arguments.

However, Orrick said he wasn’t convinced that allegations could survive if the artists failed to prove that the generated images were substantially similar to their works.

more from across site and SHARED ros bottom lb

More from across our site

Jinwon Chun discusses the need for vigilance, his love for iced coffee, and preparing for INTA
Karl Barnfather’s new patent practice will focus on protecting and enforcing tech innovations in the electronics, AI, and software industries
Partner Ranjini Acharya explains how her Federal Circuit debut resulted in her convincing the court to rule that machine learning technology was not patent-eligible
Paul Hastings and Smart & Biggar also won multiple awards, while Baker McKenzie picked up a significant prize
Burford Capital study finds that in-house lawyers have become more likely to monetise patents, but that their IP portfolios are still underutilised
Robert Reading and Faidon Zisis at Clarivate unpick some of the data surrounding music-related trademarks
China's latest IP litigation statistics and a high-profile hire by O'Melveny were also among the top talking points this week
David Aylen, who spent more than 20 years at Gowling WLG, has joined United Trademark and Patent Services as of counsel in the UAE
Europe is among the most lucrative legal markets for PE firms to bet on, but clients’ reactions will decide whether external investment drives success
Rulings of note covered pre-June 2023 infringements and jurisdiction over non-UPC states, while winners of Managing IP’s EMEA Awards acted in multiple cases
Gift this article