According to Engadget, Midjourney filed a discovery request demanding that Disney, Universal, and Warner Bros. Discovery reveal internal documents on their AI use—including business plans, research datasets, and model architecture details. The AI image generator is defending itself against a copyright lawsuit alleging it trained on protected images of Superman, Batman, and other studio-owned characters. Midjourney's argument: the studios use the same training methods they're accusing it of.
This transforms the case from a narrow copyright question into a symmetry trap. If the court grants broad discovery, judges will be examining whether the plaintiffs themselves crossed the legal lines they want enforced against Midjourney. That precedent ripples across every AI lawsuit filed in 2026. Defendants now have a template for forcing tech giants to justify their own training practices.
The studios argue that their internal AI research is proprietary and irrelevant to whether Midjourney infringed their copyrights. They contend that discovery should stay narrowly focused on Midjourney's conduct, not become a fishing expedition into Hollywood's AI operations.
Midjourney is calling Hollywood's bluff. For two years, studios have lectured the industry about AI theft while building their own generative tools behind closed doors. Now discovery forces them to choose: either admit they use similar training methods and lose the moral high ground, or claim their AI practices are fundamentally different and explain why in court documents. The fight isn't over copyright. It's over whether big media gets to write the rules for AI while exempting itself from them.
Filed to the Technology desk · Jul 4