Elon Musk Seemingly Admits xAI Has Used OpenAI's Models to Train Its Own

While testifying connected Thursday successful national court, Elon Musk seemed to bespeak that his AI laboratory whitethorn person utilized OpenAI’s models to bid xAI’s own. He touched upon the taxable portion sitting connected the witnesser basal answering cross-examination questions from an OpenAI lawyer amid his ongoing ineligible conflict against the ChatGPT-maker.

This is the exchange, arsenic champion arsenic WIRED could seizure it:

OpenAI Lawyer William Savitt: Do you cognize what distillation is?

Musk: It means to usage 1 AI exemplary to bid different AI model.

Savitt: Has xAI done that with OpenAI?

Musk: Generally each the AI companies [do that].

Savitt: So that’s a yes.

Musk: Partly.

Distillation is simply a method wherever a smaller AI exemplary is trained to mimic the behaviour of a larger, much susceptible model, making it cheaper and faster to tally portion preserving overmuch of its performance.

OpenAI’s lawyer, William Savitt, past asked whether OpenAI’s exertion had been utilized successful immoderate mode to make xAI.

Savitt: Has OpenAI exertion been utilized successful immoderate mode successful to make xAI?

Musk: It is modular signifier to usage different AI s to validate your AI.

OpenAI and xAI did not instantly respond to WIRED’s petition for comment.

OpenAI has been trying to forestall its competitors from distilling its AI models, successful particular, the Chinese AI laboratory DeepSeek. In a February 2026 memo to a House committee, OpenAI wrote that it has “taken steps to support and harden our models against distillation.” In that memo, OpenAI said it was focused connected ensuring a playing tract successful which “China can’t beforehand autocratic AI by appropriating and repackaging American innovation.”

The Trump medication has besides taken steps to forestall Chinese companies from distilling American AI models. Michael Kratsios, the White House’s manager of the bureau of subject and exertion policy, said successful an April 2026 memo that it would stock accusation with US AI companies astir overseas distillation. Kratsios said successful a station connected X that the “U.S. authorities is committed to the escaped and just improvement of AI technologies crossed a competitory ecosystem.”

American AI labs person utilized each other's AI models successful different ways, to trial advancement and measure safety. But successful today’s competitory landscape, immoderate AI companies person wholly chopped disconnected rival labs. In August 2025, Anthropic blocked OpenAI’s entree to its Claude coding models aft the institution alleged that its presumption of work had been violated. More recently, Anthropic chopped disconnected xAI from utilizing its AI models for coding arsenic well.

In his multi-day cross-examination of Musk, Savitt has questioned Musk astir his attempts to presume power of OpenAI, and subsequently, his quest to bushed the ChatGPT-maker. On Wednesday, Savitt presented emails and texts from 2017 to enactment a enactment of questions arsenic to whether Musk squeezed OpenAI by withholding backing and hiring distant cardinal researchers.