Elon Musk Appears to Confirm xAI Used OpenAI Models for Training

Elon Musk, the visionary entrepreneur behind xAI, appeared in federal court on Thursday, where he seemed to acknowledge that his artificial intelligence lab might have leveraged OpenAI's models to train its own. This revelation emerged during his cross-examination by an OpenAI attorney, William Savitt, as part of Musk’s ongoing legal dispute against the ChatGPT developer. The exchange centered on the controversial practice of "distillation," a technique increasingly scrutinized in the competitive AI landscape. During the intense questioning, Savitt directly asked Musk about distillation, to which Musk defined it as "to use one AI model to train another AI model." When pressed on whether xAI had engaged in this practice with OpenAI, Musk initially responded, "Generally all the AI companies [do that]," before conceding, "Partly." Savitt further inquired if OpenAI's technology had been used in any way to develop xAI, to which Musk replied it is "standard practice to use other AIs to validate your AI." Distillation itself is a method where a smaller AI model learns to emulate the behavior of a larger, more sophisticated one, aiming for cost-efficiency and faster operation while retaining significant performance. OpenAI, for its part, has been actively working to safeguard its proprietary models from such distillation, particularly from foreign competitors like the Chinese AI lab DeepSeek. A February 2026 memo from OpenAI to a House committee highlighted their proactive "steps to protect and harden our models against distillation." The company emphasized its commitment to fostering a level playing field, specifically to prevent "China can’t advance autocratic AI by appropriating and repackaging American innovation." This concern is not limited to private companies; the US government has also intervened. In April 2026, the Trump administration, through Michael Kratsios, then director of the White House’s office of science and technology policy, issued a memo outlining plans to share intelligence with US AI companies regarding foreign distillation efforts. Kratsios publicly stated on X that the "US government is committed to the free and fair development of AI technologies across a competitive ecosystem," underscoring the strategic importance of protecting domestic AI advancements. While American AI labs have historically collaborated, using each other's models for testing progress and assessing safety, the current fiercely competitive environment has led some to sever ties. For instance, in August 2025, Anthropic blocked OpenAI's access to its Claude coding models, citing alleged violations of its terms of service. More recently, Anthropic extended this restriction to xAI, preventing it from utilizing its AI models for coding tasks. These incidents underscore the escalating tensions and proprietary concerns within the rapidly evolving artificial intelligence sector, further complicated by Savitt's broader cross-examination, which has delved into Musk's past attempts to control OpenAI and his subsequent ambition to outcompete the company.