When three of the biggest studios in Hollywood team up to sue an AI company, you’d expect a clean-cut story. Big corporations protecting their intellectual property from a startup that lets anyone generate a photorealistic Bart Simpson or Darth Vader with a text prompt. Open and shut, right?
Not quite.
Here’s the twist that makes this case fascinating: Midjourney is now asking the court to force Disney, Universal, and Warner Bros. to reveal how *they* use AI behind closed doors. And honestly? That request feels a lot more reasonable than the studios want to admit.
The Battle So Far
Disney and Universal filed suit against Midjourney in June 2025, alleging the AI image generator enables users to create images of copyrighted characters. Warner Bros. joined in September with its own roster — Batman, Superman, the whole Justice League. The studios argue that Midjourney’s training on their content is copyright infringement, not fair use, and they’re seeking injunctions and statutory damages that could go up to $150,000 per work.
Standard stuff in the era of generative AI. We’ve seen similar lawsuits from Getty Images, from authors, from artists. The playbook is familiar.
But here’s where this case breaks from the script.
What Midjourney Is Asking For
Midjourney’s defense is built on fair use — the legal doctrine that allows limited use of copyrighted material without permission for purposes like criticism, research, and education. To bolster that defense, their legal team filed a motion asking the court to compel the studios to hand over documents about their own internal AI projects.
Specifically, they want:
- All documents about the studios’ generative AI business strategies and training datasets
- Every prompt their employees have entered into Midjourney’s own service
- All resulting outputs the studios generated
- Internal communications about AI model training, development, and usage
The argument is sharp: if Disney, Universal, and Warner Bros. are training their own AI models on unlicensed content for storyboarding, ideation, or production — the exact same behavior they’re suing over — then they’re essentially saying “it’s infringement when *you* do it, but innovation when *we* do it.”
Midjourney’s filing puts it bluntly: the studios are “cherry-picking” documents that support their market harm claims while withholding evidence that would reveal their own AI practices. The company argues that internal AI usage by the studios would demonstrate an “industry custom” of training on unlicensed copyrighted content.
The Studios’ Response
Unsurprisingly, the studios called it a fishing expedition. Their lead attorney, David Singer, said the studios “simply want Midjourney to stop copying their movies and TV shows” — not to shut down the company entirely.
A judge has already limited discovery to “consumer-facing” AI products, meaning the studios only need to disclose AI usage that led to public-facing content. Internal projects were explicitly excluded. Midjourney is now trying to overturn that limitation.
And I get why the studios are resisting. Being forced to open up your internal R&D in the middle of a lawsuit you initiated? That’s every corporate legal team’s nightmare. But it also feels like the natural consequence of suing someone for behavior you might be engaging in yourself.
Why This Case Matters Beyond Hollywood
This isn’t just a Hollywood spat. The outcome of this discovery battle could set a precedent for how fair use applies to AI training across the entire industry. There are dozens of similar lawsuits making their way through the courts — authors against OpenAI, artists against Stability AI, photographers against every image generator on the planet.
If the court forces studios to disclose their AI practices, it could create a wave of transparency requirements for any company suing an AI developer. Every plaintiff would have to answer the question: “Are you doing the same thing you’re accusing us of?”
That’s uncomfortable for a lot of big players. And when you consider the sheer scale of investment flowing into AI — Microsoft just dropped $2.5 billion on an AI deployment company, Meta is spending up to $145 billion on AI infrastructure — the legal uncertainty around fair use starts looking like a multi-trillion-dollar question mark.
Think about it like a boxing match — you don’t get to throw punches and then complain when your opponent throws one back. Midjourney isn’t asking for anything unreasonable. They’re saying, “If you want to prove that our training on your content caused real harm, show us your own AI strategy. If you’re doing the same thing internally, the harm argument falls apart.”
The Bigger Question Nobody’s Asking
Beyond the legal maneuvering, this case surfaces a question we haven’t really grappled with as a society: who gets to decide what “fair use” means in the age of AI?
The current answer, as with most tech-legal questions, is “whoever has the most expensive lawyers.” Disney, Universal, and Warner Bros. are three of the most powerful entertainment companies on earth. They have bottomless legal budgets, decades of IP protection experience, and the full weight of Hollywood behind them.
Midjourney is a startup. A well-funded one, sure — but a startup nonetheless.
But Midjourney’s countersuit strategy is clever. Instead of just defending fair use (which is an uphill battle against the studio machine), they’re turning the spotlight back on the accusers. It’s a chess move, not a boxing move. You don’t meet force with force — you change the board entirely.
As someone who works in tech and follows AI closely, I find this approach refreshing. The AI industry has spent two years playing defense — apologizing for training data, implementing opt-outs, negotiating licensing deals. Midjourney is trying something different: saying “prove you’re clean before you call me dirty.”
Where This Is Headed
The discovery hearing is likely weeks, not months, away. Given the judge’s previous ruling limiting discovery to consumer-facing products, Midjourney faces an uphill battle. But the public argument has already been won — the headlines are “Midjourney demands studios reveal AI usage,” not “Studios fight for IP protection.”
That narrative shift matters. In the court of public opinion, the idea that Hollywood studios might be doing exactly what they’re suing over is too juicy to ignore. And in a legal landscape where public perception can influence everything from jury selection to legislative action, that shift could have real consequences.
The studios’ lawsuit against Hailuo — another AI video generator — suggests they’re not backing down. They’re going after every AI company that touches their IP. But Midjourney’s countermove has at least forced them to answer uncomfortable questions.
Bottom line: this case is going to shape how AI companies train their models, how studios develop their AI strategies, and how courts interpret fair use in the generative AI era. It’s not just a lawsuit — it’s a referendum on who gets to benefit from AI, and on what terms.
And if Midjourney succeeds in forcing transparency, every company with an AI side project should be paying close attention. Because the answer to “what’s your AI strategy?” is about to become a lot harder to dodge.
Like I covered in my look at the AI hype cycle and Zuckerberg’s agent reality check, the gap between AI hype and real-world deployment is getting harder to ignore. The Midjourney case is another chapter in that same story — everyone wants AI’s benefits, but nobody wants to be seen as the one cutting corners.
Where does that leave us? Watching, waiting, and hoping the courts get this one right. Because the legal framework we build today will define AI innovation for the next decade.
If you’re interested in how AI companies are handling the legal landscape, I also wrote about Cloudflare’s move to make AI companies pay for publisher content. These pieces — the legal battles, the infrastructure shifts, the corporate maneuvering — are all part of the same puzzle. We’re building the rules of AI governance in real time, one court case and corporate policy at a time.