The legal fight between Midjourney and three of Hollywood's biggest studios has entered its most contentious phase yet, with the AI image generator asking a federal court to force Disney, Universal, and Warner Bros. Discovery to hand over detailed records of their own internal AI use. The demand marks an aggressive turn in a case that began as a straightforward copyright dispute and is increasingly becoming a test of how much transparency courts can compel from both sides of the generative AI debate.
Original lawsuits against Midjourney were filed by Disney and Universal in June 2025, claiming that their image generation models would generate unauthorized renditions of the copyrighted characters like Bart Simpson and Darth Vader. Warner Bros. Discovery filed its lawsuit in September, including a set of copyrighted characters like Batman and Superman.
The studios claimed that there should be an injunction for this infringement along with statutory damages that may run up to $150,000 per item in the case of intentional violation. Midjourney has always claimed that using public images of copyrighted characters for training their models comes under fair use.
The lawsuit has entered the discovery phase, which involves the exchange of evidence by both parties, and Midjourney is seeking much wider access than was initially granted by a court ruling in June 2026. This decision restricted the disclosure obligations of the studios to the use of AI only in relation to consumer products.
Midjourney is now asking the court to overturn that limit entirely, seeking internal documents covering the studios' AI business strategies, training datasets, model weights, and board-level presentations. It also requests every prompt that studio employees have entered into Midjourney's platform, not just the specific prompts related to the allegedly infringing images central to the case.
Midjourney's underlying argument rests on what's known as an "unclean hands" defense: if the studios are using generative AI internally for tasks like storyboarding or early-stage concept work without properly licensing every image involved, that could suggest the industry itself already treats this kind of AI training as standard, unlicensed practice. In court filings, Midjourney has stated that this evidence could show the studios are engaging in the very conduct they're suing over.
The studios' lead attorney, David Singer, dismissed the request as a "fishing expedition," maintaining that the case is narrowly about stopping unauthorized copying and distribution of copyrighted characters, not about halting AI development or shutting Midjourney down.
Also Read: YouTube AI Copyright Case: Apple Seeks Dismissal of AI Lawsuit Over YouTube Video Training Claims
Legal observers are watching the outcome closely, since whichever way the court rules on what evidence can be admitted could set a precedent affecting numerous other AI copyright cases working through the courts. If Midjourney succeeds in obtaining the studios' internal AI records, it could reshape how courts weigh fair use and industry-custom arguments in future disputes between AI companies and content owners.
For now, the case remains a pointed illustration of a central tension running through the generative AI industry: media companies suing over unauthorized AI training while facing questions about their own internal AI practices.