Midjourney Strikes Back: “Aren’t the Studios Using AI, Too?”

The copyright battle between Hollywood’s major studios and AI image company Midjourney has changed direction. Midjourney, the defendant, is now demanding that Disney, Universal and Warner Bros. reveal in court how they use artificial intelligence inside their own operations. The studios that accused Midjourney of infringement now find themselves defending their own AI practices.

There is a structural reason this fight keeps escalating. Generative AI has already seeped into nearly every stage of content production and distribution, from storyboarding and visual effects to marketing, and the studios themselves have been building or adopting AI tools internally. Many of those tools are suspected of resting on the same foundation as Midjourney’s: models trained on large volumes of web data.

As long as the core allegation is unauthorized use of copyrighted works for AI training, whether the plaintiffs train and deploy AI in the same manner becomes a variable that could decide the case. That is precisely where Midjourney is digging.

Case Status: A First Ruling Limited to Consumer-Facing AI

According to Variety, Disney and Universal sued Midjourney last year, accusing the AI image lab of enabling massive infringement of their copyrighted characters, and Warner Bros. joined as a plaintiff. Midjourney has asserted a fair use defense while arguing that the studios engage in the very same AI practices.

On June 15, Magistrate Judge Joel Richlin denied Midjourney’s attempt to obtain broad information about the studios’ AI use, finding it irrelevant to whether Midjourney infringed the studios’ copyrights. The studios were required to turn over information only about consumer-facing AI applications; internal AI tools were excluded from disclosure.

Midjourney’s Appeal: Training Datasets, Model Weights, Even Board Materials

Midjourney has now asked District Judge John Kronstadt to overturn that decision, filing a motion this week. The scope of what it seeks is wide: the studios’ AI business plans, research reports, training datasets, model weights, and board meeting presentations on AI — materials that would show how AI tools are used to create and market movies and television shows.

In the motion, Midjourney’s attorney Bobby Ghajar argued that if the 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.

He further contended that if the studios are developing image-generating AI models trained on unlicensed third-party copyrighted data for internal uses such as storyboarding or ideating film and TV content, that would demonstrate an industry custom — even among the studios themselves — of downloading and training AI on unlicensed copyrighted material. The strategy is to establish unauthorized training as standard industry practice and thereby tilt the fair use analysis in Midjourney’s favor.

The Studios’ Rebuttal: A “Fishing Expedition”

The studios’ lead attorney, David Singer, has countered that Midjourney is mounting a fishing expedition designed to distract from its own misconduct. In an earlier filing opposing Midjourney’s discovery motion, he stressed that the plaintiffs are not trying to stop AI technology or shut down Midjourney’s business; they simply want Midjourney to stop copying their movies and TV shows and to stop distributing, displaying, performing and creating derivative works built on their famous characters without authorization — the same rights any copyright holder would assert against any infringer.

What It Means: Hollywood’s Internal AI, Summoned to Court

The stakes reach beyond the fate of a single company. If Judge Kronstadt reverses the magistrate’s ruling and brings internal AI materials within the scope of discovery, the record will show — for the first time in a courtroom — what data, what models and what stages of production Hollywood’s studios rely on. That evidence could ripple into other pending AI copyright cases, as well as labor negotiations and regulatory debates over AI in content production.

If, on the other hand, disclosure stays limited to consumer-facing AI, the studios preserve their current strategy of separating internal AI adoption from a hard external line on copyright. The content industry as a whole now stands on a dual structure — using AI as a tool while refusing to legitimize it as the other side’s weapon — and this case has become the first test of whether that structure can hold up in court.

Source: Variety, “Midjourney Seeks to Reveal Studios’ Use of AI in High-Stakes Copyright Battle” (Gene Maddaus)