It sure looks like OpenAI trained Sora on game content — and legal experts say that could be a problem | TechCrunch
It seems that OpenAI trained its video-generating model, Sora, on copyrighted video game content. And that could be legally problematic.