MiniMax loses key motion in Disney copyright fight over Hailuo generative AI

The MiniMax decision is the latest sign that courts are increasingly willing to let high‑stakes copyright cases against generative AI providers go the distance. By refusing to toss Disney and its peers’ lawsuit at an early stage, the judge is effectively saying that arguments over fair use, training data, and model outputs deserve full discovery and trial, not a quick dismissal. That’s bad news for any model builder whose datasets or outputs lean heavily on unlicensed entertainment IP.([economictimes.indiatimes.com](https://economictimes.indiatimes.com/tech/technology/chinas-minimax-loses-bid-to-end-disney-copyright-lawsuit-over-ai-system/articleshow/131332455.cms/?utm_source=openai))

For the AGI race, the biggest risk is not this single Chinese startup but the precedent it may set. If US courts ultimately find that training on or reproducing copyrighted audiovisual material at scale is infringing, it could force a pivot toward more tightly licensed or synthetic corpora for multimodal frontier models. That would raise costs, slow iteration, and advantage players with deep content libraries or strong licensing channels. It also adds legal overhang for any global lab trying to serve US customers with image‑ or video‑generation systems that were trained abroad under looser regimes.