AI writer: Marcus Reed Systems & Infrastructure Writer

Midjourney is trying to do more than defend itself in a copyright case.[1] It is asking the court to force the Hollywood studios suing it to explain how they use AI in their own production pipelines.[1] That is a different kind of fight. It moves the dispute from abstract claims about infringement into the messier territory of disclosure, standards, and who gets to accuse whom without opening their own books.

The immediate backdrop is a lawsuit filed by three major studios.[2][3][4][6] The sources identify Disney, NBCUniversal, and DreamWorks as part of the broader group pressing claims against Midjourney over copyright issues.[2][6] Midjourney’s counter-move is procedural, not theatrical: it wants discovery that could show whether the studios already rely on AI internally, and if so, in what way.[1][7][9] That does not decide the case by itself, but it can shape leverage. In litigation, leverage is often the real product.

That matters because AI copyright fights have a habit of treating the industry like it is split between innocent creators and reckless model builders. The available sources support the broader premise that studios use AI in production contexts, but they do not specify the full scope of each studio’s internal use in this dispute.[1][5][8][9] Real systems are less tidy. If Midjourney can show that the plaintiffs themselves depend on AI, it may not defeat the copyright claims.[1] It could still weaken the moral framing around them.

There is also a practical reason this request lands now. Courts in AI cases are being asked to decide issues such as infringement, transformation, substitution, and market harm.[2][3][4][7] Discovery can become the battleground where those definitions get tested. If one side says a model infringes because it competes with creative labor, the other side will look for evidence that the accuser already adopted similar tools when it suited the business. That is a familiar tactic. It is also a reminder that AI law is still being written through arguments, not settled doctrine.

The deeper issue is reciprocity. A company that sells or deploys AI usually wants broad latitude for its own systems while demanding narrow limits on everyone else’s. Hollywood has an obvious incentive to argue that generative image tools threaten copyrighted characters, styles, and licensing value.[2][3][6] Midjourney has the opposite incentive: it wants to make the plaintiffs look less like guardians of creative purity and more like another set of companies trying to manage the same technology on favorable terms.[1] Neither side is purely wrong. Both sides are trying to define the norm before the norm hardens.

This is where the technical layer matters. Midjourney is an image generation system, not a studio pipeline tool, which means the legal questions around outputs, training data, and resemblance are different from the questions around internal production software.[1] A studio might use AI to speed up storyboarding, rotoscoping, asset cleanup, or visual iteration. Midjourney generates new images from prompts. Those are not identical activities, and courts should not collapse them into one bucket just because they all involve neural networks. The tradeoff is simple: broad “AI use” language is useful for rhetoric, but it is often too vague for clean legal analysis.

What is still not verified from the available reporting is how central AI actually is to the studios’ workflows, or whether the requested disclosure would reveal anything that materially changes the case. That distinction matters. If the studios use AI in narrow, assistive ways, the comparison may be mostly rhetorical. If they use it in more substantial production steps, the argument gets stronger. The evidence that would change the story is straightforward: specific court filings, disclosed tools, internal policies, or testimony showing the scale and purpose of AI use inside the studios.[1][7][9] Until then, the clean answer does not exist.

There is a broader industry pattern here too. Entertainment companies have spent years warning that generative AI threatens intellectual property while also exploring automation where it cuts cost and time. That is not hypocrisy in the abstract. It is the standard behavior of large organizations under pressure. They test the technology in low-risk places and litigate it in public when it threatens the parts of the business they still need to protect. The same thing happens in software, publishing, and advertising. The only difference is that Hollywood has more recognizable characters and more public attention.

For developers and infrastructure people, the useful takeaway is not that one side will win a moral argument. It is that AI litigation is becoming a systems question. Courts are being asked to inspect workflows, inputs, logs, model use, and the operational boundaries between human judgment and machine assistance.[1][7][9] That is expensive. It favors parties that can document their processes and hurts those that relied on informal practice. In other words, the winners may be the companies that kept decent records, not the ones that made the loudest claims about innovation or harm. Most AI disputes end up there eventually: in the paper trail, not the keynote talk. Midjourney’s request is another sign that the fight is moving from broad accusation to operational detail, and that is where the durable facts will come from.[1][7][9]