Midjourney Demands Studios Disclose Their Own AI Training Practices
The image-generation startup wants discovery to include how Disney, Universal, and Warner Bros. use copyrighted content internally for AI development.
Midjourney pushes for broader discovery in copyright case
Midjourney is asking a court to force Disney, Universal, and Warner Bros. to disclose comprehensive details about their own artificial intelligence practices, including internal uses that never reach consumers, according to a new legal filing in the ongoing copyright dispute.
The three studios sued Midjourney over alleged copyright infringement, claiming the startup's image-generation models can produce images of their protected characters including Bart Simpson and Darth Vader. Midjourney maintains that training AI models on copyrighted images falls under fair use protections.
A judge previously ruled the studios must provide information about their generative AI usage during discovery, but limited that requirement to "consumer-facing" videos and images. Midjourney's latest motion seeks to eliminate that restriction.
The core argument: revealing potential hypocrisy
Midjourney argues the current limitation allows studios to "cherry-pick only those documents they believe support their market harm claims" while withholding evidence that could support the startup's defense. The company claims the restricted documents could reveal whether studios are "doing exactly what they are suing Midjourney for doing" behind closed doors.
Specifically, Midjourney contends that if the studios are developing image-generating AI models for internal purposes like storyboarding or content ideation, that would demonstrate "it is an industry custom, even among the studios themselves, to download and train AI on unlicensed copyrighted content."
The startup also wants access to all prompts the studios have used in Midjourney and their resulting outputs, not just those that allegedly produced infringing images.
Studios push back on discovery scope
David Singer, lead attorney for the studios, has characterized Midjourney's discovery requests as a "fishing expedition." He emphasized that the studios don't aim to halt AI technology or shut down Midjourney's business entirely. Rather, they want the company to "stop copying their movies and TV shows" and cease distributing, displaying, performing, or creating derivative works featuring their characters without authorization.
Why it matters
This discovery dispute could establish important precedent for how copyright law applies when alleged infringers argue that plaintiffs engage in similar practices. If Midjourney succeeds in obtaining evidence that major studios train AI models on unlicensed copyrighted content for internal use, it could complicate the studios' claims of market harm and potentially support fair use arguments across the industry. The outcome may influence how courts balance transparency requirements against corporate confidentiality in AI-related copyright litigation.
These details were first reported by TechCrunch.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call
