Owners, creators and developers are watching these cases closely for much-needed precedent to guide AI training on original works.
Warner Bros. Discovery and affiliates sued Midjourney, Inc., in the Central District of California on Sept. 5, alleging direct and secondary infringement under the Copyright Act in its use of copyrighted works to train Midjourney’s AI design models (Warner Bros. v. Midjourney, No. 2:25-cv-08376, C.D. Calif.). The Warner Bros. litigation follows that initiated by Disney Enterprises and Universal Studios which filed a joint suit in the same court in June.
Here are the chief allegations as detailed in Warner’s complaint:
Unauthorized Use: Warner alleges Midjourney copied and used thousands of its copyrighted works—including characters like Superman, Batman, Wonder Woman, Bugs Bunny, and Scooby-Doo—without permission.
Infringing Outputs: Midjourney enables users to generate, display, and download images and videos closely resembling Warner’s iconic characters and scenes, providing specific examples in the complaint.
Knowledge and Willfulness: The complaint claims Midjourney knew of its infringement, citing instances where the company temporarily blocked video generation for certain characters but later removed those protections. Warner describes this conduct as “systematic, ongoing, and willful.”
Lack of Protections: Warner asserts that, unlike other platforms, Midjourney does not implement meaningful copyright safeguards—such as excluding protected works from training, rejecting infringing prompts, or screening outputs.
Commercial Benefit and Market Harm: Warner claims Midjourney’s actions are profit-driven, generating hundreds of millions in annual revenue while diverting consumers from licensed products and causing confusion about the legality of AI outputs.
Secondary Liability: Warner pleads that, if users are found to be direct infringers, Midjourney is vicariously and contributorily liable for facilitating and encouraging infringement.
Warner seeks actual and statutory damages (up to $150,000 per work), disgorgement of profits, and injunctive relief to stop further infringement.
In their June 11, complaint, Disney and Universal Studios likewise allege Midjourney’s AI service was trained on, and outputs, unauthorized derivatives of characters from their libraries—including Star Wars, Marvel, The Simpsons, Pixar, Shrek, and Minions. Midjourney acts like a “virtual vending machine” for infringing copies. It ignores cease-and-desist letters and fails to block copyright violations while restricting other content types, the suit maintains. The studios offer side-by-side comparisons and assert that Midjourney reproduces protected characters even absent explicit prompts (Disney Enterprises Inc. et al v. Midjourney Inc., 2:25-cv-05275, C.D.Calif.).
“Highly Transformative”
Midjourney denies nearly all allegations, asserting its models are trained on billions of public images in a “highly transformative” process akin to human artistic learning. The company invokes fair use, citing recent federal rulings, and claims its platform enables user expression rather than infringing copies. Midjourney argues that alleged violations should have been handled via Digital Millenium Copyright Act (DMCA) notice-and-takedown, which the studios bypassed.
Affirmative defenses include fair use, lack of standing, licensing, acquiescence, estoppel, waiver, unclean hands, de minimis copying, and DMCA safe harbor. Midjourney requests dismissal and a jury trial.
On Sept. 5, Disney and Universal notified the court that their suit against Midjourney is related to the Warner Bros. Discovery action and asked that they be consolidated. All of the studios are represented by Jenner & Block.
Midjourney is also a defendant in a case in the Northern District of California. A group of artists sued the company, along with Stability AI and DeviantArt for using their artwork – billions of images — in training datasets, citing copyright infringement, Digital Millenium Copyright Act (DMCA) violations, and false endorsement. This visual generative AI case is set for trial in 2027 (Sarah Andersen, et al. v. Stability AI Ltd., et al., No. 3:23-cv-00201-WHO, N.D.Calif.). Read about other cases.
Conclusion and Key Takeaways
The ongoing lawsuits against Midjourney by major studios like Warner Bros., Disney, and Universal, along with the artist-led case in Northern California, cast a spotlight on the tension between copyright law and AI-generated content. Key takeaways include the importance of clear copyright safeguards, the legal complexities of training AI on protected works, and the significant commercial and market impact these technologies can have on rights holders.
The outcomes of these cases are certain to set nationwide precedent on the use of copyrighted works to train AI large language models. Creators, intellectual property owners, and AI developers should closely monitor these developments, as the resulting decisions will shape the future boundaries of innovation and copyright protection in the digital age.
