Midjourney, a company that makes an AI art generator, has asked a court to force Disney, Universal, and Warner Bros. to reveal how they use artificial intelligence internally. The studios are suing Midjourney for copyright infringement; Midjourney's request would expose whether the studios use similar AI technology on their own copyrighted characters.

The court filing happened on March 28, 2026, in federal court in California. Disney and Universal sued Midjourney together in June 2025 for allegedly training its AI on their characters without permission. Warner Bros. sued separately in September. The two cases are now being handled together.

What Midjourney Wants to Know

A judge previously limited what Midjourney could ask the studios to hand over — only AI tools they had actually used to make public content (images or videos people could see). According to TechCrunch on July 4, 2026, Midjourney now wants to expand that. It wants the studios to also reveal behind-the-scenes AI work — tools used in production that touched their characters but never got released to the public.

Midjourney's main defense is that using AI trained on copyrighted material is actually legal. It is called "fair use," a legal concept that allows certain uses of copyrighted work without permission. Midjourney thinks it has a stronger argument if it can show that Disney, Universal, and Warner Bros. use similar AI tools themselves. If the studios are already doing what they say Midjourney shouldn't be doing, the law becomes harder to apply one way.

The studios' lawyers called this approach a "fishing expedition," according to Variety — meaning they think Midjourney is just searching randomly for anything that might help its case. The judge who set the original limit seemed to agree the studios shouldn't have to disclose everything. Now Midjourney is asking him to change his mind.

Why This Matters

If the judge agrees with Midjourney, the three studios would have to release confidential information about how they use AI. They rely on AI for visual effects, sound editing, marketing, and post-production work — but they do not usually say exactly which systems they use or how. Being forced to do so would be unprecedented.

This is not just a legal maneuver. If Midjourney can prove the studios themselves use the same type of AI on their own characters, that makes Midjourney's "fair use" argument sound more reasonable to a jury. Whether a judge will let that evidence be used at trial is a separate question, but discovery — the part where both sides share information — is how you get evidence into the case.

The studios face a logical problem of their own. They say Midjourney hurt them by training AI on their characters. But if the studios are using AI on those same characters inside their own companies without paying the people who created them — writers, artists, actors — then the question of who owes what becomes much less clear. No court has yet said what "fair use" really means for AI trained on huge amounts of copyrighted work, and what the judge learns from the studios' own practices could shape that answer.

The judge will decide soon. His previous order only required the studios to reveal AI that made public content. Midjourney is asking him to require them to reveal everything — including work that never left the building. The decision will set rules for how much AI companies can investigate their accusers' technology practices in future lawsuits.