Midjourney wants to expose how Disney, Universal, and Warner Bros. are using AI
The studios are suing Midjourney for letting users mass-generate their copyrighted characters. Now Midjourney is firing back with a clever legal move: forcing the studios to reveal how they secretly use AI themselves. Here’s the high-stakes courtroom battle, and why both sides have a point.
The war between Hollywood and AI just took a fascinating turn. Midjourney, the AI image generator being sued by the biggest studios in the business, is trying to flip the script, by forcing those studios to reveal how they use AI behind closed doors.
It’s a bold legal counterpunch in a case that could reshape the entire AI industry. Here’s what’s happening, what each side is arguing, and why this fight matters far beyond one lawsuit.
First, the lawsuit itself
Let’s back up and explain how we got here.
Starting in June 2025, Disney and Universal, later joined by Warner Bros. Discovery, sued Midjourney for copyright infringement. Their core accusation: Midjourney’s tool lets paying subscribers generate endless images of the studios’ famous copyrighted characters, everyone from Darth Vader and Shrek to Batman, Superman, and Homer Simpson, and that Midjourney trained its AI on their intellectual property without permission.
Disney’s complaint memorably called Midjourney a “bottomless pit of plagiarism.” The studios argue that a subscriber can type a simple prompt like “Batman” and get a downloadable, on-model image they could slap on a T-shirt, directly competing with the studios’ own licensed merchandise.
The stakes are enormous. The studios are seeking up to $150,000 per infringed work. With potentially millions of infringing images, that math could add up to a company-ending sum for Midjourney.
Midjourney’s clever counter-move
Here’s the twist that makes this so interesting.
Rather than just playing defense, Midjourney is going on the offensive during the case’s “discovery” phase, the part of a lawsuit where each side must hand over relevant evidence. Midjourney is demanding that Disney, Universal, and Warner Bros. reveal how they use AI internally, including their AI training datasets, research reports, model data, and even board meeting presentations about their AI plans.
Why? Because Midjourney’s entire strategy hinges on a “you do it too” argument. Its defense rests on two ideas:
Fair use: the legal principle that some uses of copyrighted material are permitted.
“Unclean hands”: a legal doctrine that basically says you can’t punish someone for doing the same thing you’re doing yourself.
Midjourney’s lawyers argue that if the studios are also training their own AI models on unlicensed copyrighted material, to storyboard scenes or brainstorm content, then that AI training is simply an “industry custom.” As one filing put it: “If Plaintiffs are doing the very thing they seek to punish, that evidence goes to the heart of Midjourney’s fair use and unclean hands defenses.”
The judge’s ruling (so far)
Here’s where the case actually stands.
In mid-June 2026, a magistrate judge partly shut Midjourney down. The judge ruled that the studios only have to reveal information about their consumer-facing AI tools, the ones the public uses, not their secret, internal AI work. The judge found Midjourney’s broader demands largely irrelevant to the actual question of whether Midjourney infringed on the studios’ copyrights.
Midjourney isn’t accepting that. This week, its lawyers filed a motion asking the district judge, John Kronstadt, to overturn the magistrate’s decision and force the studios to open up about their behind-the-scenes AI use. That ruling is now pending.
What each side is arguing
Here’s the fair breakdown, because both sides have genuinely strong points.
The studios’ case is straightforward and powerful: Midjourney is running a business built on reproducing their characters for profit, plain and simple. Their lead attorney, David Singer, dismissed Midjourney’s discovery demands as a “fishing expedition” meant to distract from its own conduct. As Singer put it, the studios just want Midjourney “to stop copying their movies and TV shows”, the same right any copyright holder would defend. And they’re not trying to ban AI, just to stop the unauthorized use of their characters.
Midjourney’s case taps into a real hypocrisy angle: the studios are aggressively adopting AI themselves. If Disney or Universal are quietly using AI tools trained on copyrighted material to make their own films and shows, Midjourney argues, then the studios “cannot have it both ways”, profiting from the same AI practices they’re suing over. It’s a genuinely awkward question for an industry that’s simultaneously fighting AI in court and racing to use it in production.
Why this case matters way beyond Midjourney
Here’s the bigger picture, and it’s a big one.
This is the first major battle between traditional Hollywood and the AI industry, and its outcome could ripple across dozens of similar lawsuits (against OpenAI, Anthropic, Stability AI, and others). The specific discovery fight, whether an AI company can force copyright holders to expose their own AI practices, could set a powerful precedent.
If Midjourney wins that right, every AI company being sued could start demanding to see how its accusers use AI, potentially exposing an uncomfortable industry-wide truth: that everyone is training on copyrighted data. If the studios win, it reinforces limits on that strategy and keeps the focus squarely on the AI companies’ conduct.
Either way, the case is likely to help define the ground rules for how copyright works in the age of AI, rules that will affect artists, studios, tech companies, and anyone who’s ever typed a prompt.
Midjourney vs. the studios: what the AI copyright fight really comes down to
The Midjourney lawsuit has become a landmark clash, the studios say the AI firm is a plagiarism machine profiting off their characters, and Midjourney says the studios are hypocrites doing the very same AI training in secret. Right now, it’s a procedural chess match over who has to reveal what, and the judge has mostly sided with the studios, for the moment.
But the deeper question is far from settled, and it’s a genuinely hard one. Is training AI on copyrighted work theft, or is it fair use, and does it matter if the accusers are doing it too? Both sides have legitimate arguments, and honestly, no one knows yet how courts will ultimately rule. What’s certain is that this case is bigger than one AI company or one batch of cartoon characters.
It’s an early, defining skirmish in the fight to decide who owns creativity in the AI era, and everyone, from mega-studios to individual artists, has a stake in the outcome. Grab your popcorn.
This one’s going to run for a while.
Want More Clownfish TV?
This article was brought to you in part by The Reefers of more.clownfishtv.com. Free subscribers get articles like this one in their inbox. Paid subscribers get the full Clownfish TV podcast feed, livestreams, and members-only episodes that never hit YouTube.
D/REZZED is part of Clownfish TV. For more news, views, and rants on gaming, tech, and pop culture, watch @ClownfishTV on YouTube and find the podcast on Apple Podcasts, Spotify, and iHeart.
Article compiled and edited by Derek Gibbs (entertainment editor) and the Clownfish TV newsroom.
Hat Tips:
Variety (July 2026), the originating report, verified for the current discovery dispute (Midjourney’s motion to force disclosure of the studios’ internal AI use, the June magistrate ruling limiting discovery to consumer-facing AI, Judge Kronstadt’s pending review, the “if Plaintiffs are doing the very thing they seek to punish” and David Singer “fishing expedition” quotes, and the specific data sought, training datasets, model weights, board presentations)
The Hollywood Reporter and NPR (June-September 2025), verified for the underlying lawsuits (Disney/Universal in June 2025, Warner Bros. Discovery in September 2025), the “bottomless pit of plagiarism” language, the $150,000-per-work damages, the character examples (Darth Vader, Shrek, Batman, Superman, Bugs Bunny, Homer Simpson), and Midjourney’s fair-use defense
TheWrap and legal-analysis coverage (Let’s Data Science, McKool Smith) (2025-2026), verified for the fair-use and “unclean hands” defense framing, the “cannot have it both ways” argument, the case’s status as the first major Hollywood-vs-AI battle, and the precedent it could set for other AI copyright suits