Elon Musk appeared to suggest that his AI Lab may have been using OpenAI models in order to create xAI. Elon Musk mentioned the subject while on the witness stand, answering questions from an OpenAI lawyer during his ongoing federal court case. legal battle against the ChatGPT-maker.
The best WIRED could do was capture the exchange as follows:
OpenAI lawyer William Savitt What is distillation?
Musk: This is the process of using one AI model for training another AI model.
Savitt: OpenAI: Has xAI achieved this with OpenAI?
Musk: All the AI companies [do that].
Savitt: That’s an affirmative answer.
Musk: Partly.
It is possible to distill an AI model to make it more efficient and cheaper to operate, while still retaining its high performance.
William Savitt asked OpenAI’s lawyer if OpenAI’s technology was used to create xAI.
Savitt: OpenAI was used for xAI development.
Musk: Validating your AI with other AIs is a standard procedure.
WIRED’s requests for comments to OpenAI or xAI were not responded immediately.
OpenAI is trying to stop its competitors, namely the Chinese AI laboratory DeepSeek, from distilling their AI models. According to a 2026 February article memo OpenAI told a House Committee it had “taken steps to protect and harden our models against distillation.” OpenAI stated in that memo it was focused on creating a level playing field for all players. “China can’t advance autocratic AI by appropriating and repackaging American innovation.”
Trump has taken measures to keep Chinese AI companies away from American AI. Michael Kratsios is the White House director for science and technology policy. He said this in an article published on April 20, 2026. memo The company said it would provide information to US AI companies on foreign distillation. Kratsios said in a post on X This is the “US government is committed to the free and fair development of AI technologies across a competitive ecosystem.”
American AI laboratories have also used AI models from each other to assess and test safety, as well as progress. Some AI companies, in the current competitive market, have cut rival labs off. In 2025. Anthropic blocked OpenAI’s access Anthropic claimed that the terms of service were violated. Anthropic cut off xAI from using its AI models Coding is also required.
Savitt, in his multiple-day cross-examination Musk, has asked Musk questions about his attempts to assume control of OpenAI And, then, his quest for beating the ChatGPT maker. On Wednesday, Savitt provided emails and text messages from 2017 in support of a question as to whether Musk squeezed OpenAI through withholding funding or hiring key researchers.
