Midjourney, which is being sued by three of Hollywood's largest studios for copyright infringement, has changed tactics. Instead of defending itself, it wants Disney, Universal, and Warner Bros. to reveal in court how they use artificial intelligence themselves. And it suggests that behind closed doors, they are doing exactly what they are now taking Midjourney to court for.
The dispute has been ongoing since last year. Disney and Universal filed a lawsuit in mid-2025, alleging that Midjourney's image models can generate characters to which the studios own the rights. These included Bart Simpson and Darth Vader. A few months later, Warner Bros. joined the plaintiffs, objecting to images of Superman, Batman, and other characters.
Midjourney does not deny the allegations. However, it argues that training AI on copyrighted works falls under so-called fair use. And it is countering that Disney, Universal, and Warner Bros. routinely use similar generative tools internally.
Court proceedings have stalled
The current tug-of-war centers on discovery. The judge previously ruled that the studios must provide information about their use of generative AI, but only in cases where it led to specific images and videos intended for the public. The requirement did not apply to internal tools. And that is precisely what Midjourney now wants to change. In its latest filing, the company argues that such a restriction unfairly ties its hands. According to Midjourney, the studios can select only the documents that support their claims of market harm and keep the rest under wraps.
According to Midjourney, those withheld materials are the most important. The company claims they would show whether the studios are secretly doing the same thing they are suing Midjourney for. Midjourney wants each studio to turn over specific materials: training datasets, business plans, and presentations from management meetings concerning generative AI.
According to Midjourney representative Bobby Ghajar, if the plaintiffs are doing exactly what they seek to punish, that goes directly to the heart of the company's fair-use defense. Ghajar added that if the studios are developing image models trained on third-party copyrighted data for internal use, such as drawing storyboards or developing ideas for film and television, it would demonstrate that downloading unlicensed copyrighted content and using it to train AI is common practice in the industry—even among the studios themselves.
Midjourney wants to see every prompt the studios entered into Midjourney, along with the resulting images, not just those that led to the allegedly infringing outputs.
The studios push back
The other side is pushing back. The studios' lead attorney, David Singer, previously described Midjourney's attempt to obtain these materials as a fishing expedition intended to distract attention from the company's own misconduct.
Singer also emphasized that the studios are not seeking to stop AI or put Midjourney out of business. He said they simply want Midjourney to stop copying their films and television shows and to stop distributing, publicly displaying, and creating derivative works featuring their famous characters without permission. According to him, these are rights that any copyright holder would enforce against an infringer, whether AI is involved or anything else.
Of the three studios, Disney has been the most forthcoming about artificial intelligence. At the end of 2025, it announced a billion-dollar investment in OpenAI intended to bring hundreds of Disney characters to the Sora platform. The plans fell through at the beginning of this year after Sora shut down. Disney said at the time that it was not opposed to AI platforms and wanted to continue working with technologies that respect intellectual property and creators' rights.
Exactly what this cooperation will look like may now depend on how the dispute ends—and, above all, on what Midjourney ultimately manages to obtain from the three studios.
Sources: petapixel.com, gizmodo.com and variety.com
