Elon Musk has admitted in court that his company xAI uses distillation techniques on OpenAI models to train its AI, Grok. This admission comes during a trial where Musk is suing OpenAI over alleged breaches of their original nonprofit mission.

The revelation highlights the intense competition and collaboration within the AI industry as firms seek to innovate while avoiding direct copyright infringement. Distillation allows smaller players like xAI to build models that rival those of giants like Google, which may threaten the market dominance of established companies.

OpenAI has responded by joining forces with Anthropic and Google through the Frontier Model Forum to combat distillation attempts from Chinese firms. This initiative involves sharing information on how to prevent mass querying of their models. However, Musk’s testimony suggests that despite such efforts, industry practices may be more open than previously thought.

The admission has sparked debate about the legality and ethical implications of using other companies’ AI outputs in training new models. While there is no explicit legal prohibition on distillation, it may contravene terms of service agreements between firms and their users.

As this tech arms race intensifies, questions arise about transparency, competition, and the long-term development of artificial intelligence. The extent to which companies can leverage each other’s advancements could shape the future landscape of AI innovation.