The situation is getting interesting. Disney, Universal, and Warner Bros. sued Midjourney because its models can generate images of Darth Vader, Bart Simpson, and other copyrighted characters. That’s the known story. Now comes the plot twist.
‘Show Us Your AI Usage’
In new court filings from July 4th, Midjourney is demanding that the studios disclose how they themselves use AI. The allegation: the studios are developing their own image-generating AI models behind closed doors — and potentially training them on the very same unlicensed copyrighted material they’re suing Midjourney for using.
Midjourney’s argument: if the studios are doing the same thing internally, it demonstrates that training on copyrighted material is an industry practice — not piracy by a single startup.
What a Judge Has Already Ruled
A judge already ordered the studios to provide information about their own AI use — but only where it led to ‘consumer-facing’ videos and images. Midjourney is pushing further: it wants to see every prompt the studios entered into Midjourney, plus all resulting outputs — not just the prompts that produced the allegedly infringing images.
Why This Matters
This case could set a precedent for the entire AI industry. If a court finds that even the rights holders train AI on copyrighted material, it becomes very hard to prohibit that exact practice when others do it.
The studios have stayed silent on the allegations so far. But the dynamic has shifted: Midjourney is no longer just the defendant — it’s asking uncomfortable questions of its own.
Sources: