During testimony in his lawsuit against OpenAI, Elon Musk admitted under oath that his xAI startup used OpenAI's models to train Grok through model distillation. Musk argued this is standard industry practice, but the admission provides evidence central to OpenAI's defense. The trial has revealed significant tensions over proprietary model training, intellectual property, and industry norms around model improvement techniques.

What This Means for Your Business

This trial is exposing the murky IP landscape around model distillation and training practices. Organizations licensing AI models should clarify contract terms around model improvement, retraining rights, and derivative uses—courts are now actively adjudicating these disputes. The trial suggests that practices assumed to be industry-standard may face legal challenge, so review your AI licensing agreements with counsel now, before disputes emerge.