Midjourney’s Bold Gambit: Turn a Copyright Case into an AI Transparency Trial
The Midjourney copyright lawsuit is a legal battle in which major film studios accuse an AI image generator of large-scale copyright infringement, while the company responds by demanding evidence about Hollywood AI training methods and studio AI practices to support a fair use and unclean hands defense.
Midjourney is not merely defending itself; it is trying to put the entire industry’s AI habits on the record. After Disney and Universal sued in June 2025 for enabling large-scale infringement of characters like Darth Vader and Elsa, and Warner Bros. Discovery followed three months later with claims of “brazen theft” and statutory damages of USD 150,000 (approx. RM690,000) per infringed work, the usual script would cast Midjourney as the sole villain. Instead, the company argues that training on publicly available images is fair use and that the studios do the same thing with their own models. This flips the narrative: if everyone is borrowing from the same cultural commons, the question is no longer whether AI training is pure, but whether anyone can credibly claim clean hands.
The Discovery Offensive: Forcing Open the Black Box of Studio AI Practices
Midjourney’s legal strategy hinges on discovery, and it is asking for nothing less than a map of Hollywood AI training methods. Specifically, it wants AI business plans, research reports, training datasets, model weights, and even board-level presentations on AI from Disney, Warner Bros. Discovery, and Universal. That is a surgical request aimed directly at the studios’ own studio AI practices, not a vague fishing trip.
A magistrate judge recently tried to narrow this offensive, allowing the studios to withhold most internal information and requiring disclosure only for “consumer-facing” AI tools. Midjourney has now gone to the federal judge, asking to overturn that limit. The company’s argument is blunt: if the plaintiffs are training AI on copyrighted works in the same way they say Midjourney did, that evidence “goes to the heart” of its fair use and unclean hands defenses. In other words, Midjourney wants the court to decide whether the law applies equally to Silicon Valley upstarts and legacy studios, or only to one side.
Fair Use, Unclean Hands, and the Risk of Industry-Wide Hypocrisy
At the center of this AI copyright dispute is a simple, uncomfortable question: can studios sue over AI training they quietly perform themselves? Midjourney claims that training on publicly available images is fair use, and that the major studios rely on comparable methods for their own internal AI models. If discovery confirms this, the studios’ case against Midjourney weakens not only on legal grounds but on moral authority.
This is why the company leans so heavily on the “unclean hands” doctrine. It is not enough for the studios to show that Midjourney can generate images of Superman, Batman, Bugs Bunny, or other copyrighted characters; they must also show they are not “doing the very thing they seek to punish” themselves. That framing turns Hollywood’s role from injured party into potential co-participant in the same contested practices. If Midjourney succeeds, the precedent will signal that plaintiffs in AI cases cannot selectively expose others’ models while hiding their own.
From One Defendant to a Whole Industry on the Stand
The most important shift here is conceptual: this is no longer just about Midjourney’s liability but about industry-wide AI transparency. By demanding evidence about how studios use AI technologies in their shows and movies, Midjourney is pulling the conversation away from a single defendant and toward a structural question: what AI training methods are acceptable for anyone to use?
The studios’ lawyer has dismissed these discovery demands as a “fishing expedition” meant to deflect from Midjourney’s conduct. But that critique cuts both ways. If AI models are to be judged in court, courts need to see more than screenshots of outputs; they need to understand who trained what, on which data, and under which rules. The federal judge’s upcoming decision on whether to expand discovery will not only shape this case but, as reported, could set a precedent for what kind of internal AI information can be admitted in future lawsuits. Put plainly, whatever happens next will help decide whether AI copyright disputes are fought in the open, or continue as one-sided accusations backed by sealed black boxes.
