Overview
- Late June filings show Midjourney has asked a district judge to overturn a mid‑June discovery limit and compel Disney, Universal and Warner Bros. to produce internal generative‑AI records such as training datasets, model details, board presentations and prompts.
- The studios have supplied information about consumer‑facing AI projects but have refused to disclose internal tools and plans, saying the suit targets unauthorized copying of copyrighted works rather than industry experimentation.
- Midjourney argues that studio records would support its fair‑use and unclean‑hands defenses by showing whether major rights holders themselves scraped or used third‑party copyrighted material to train models.
- No decisive court ruling has been issued and lawyers say the outcome could set new limits on discovery in AI copyright cases and shape how courts weigh fair use for model training and outputs.
- The dispute comes as studios explore AI investments and platforms and as many related lawsuits move through U.S. courts, with possible consequences for creators’ rights, studio practices and transparency around model training.