Musk Appears to Admit xAI Used OpenAI Models to Train Grok

Elon Musk has seemingly acknowledged that xAI used OpenAI’s models to train its Grok chatbot, according to WIRED, amid ongoing litigation between the two AI rivals.

The reported admission comes as Musk remains in litigation against OpenAI, the company he co-founded before departing, complicating the already contentious dispute between the two companies.

Using a competitor’s AI model outputs to train a rival system — a practice sometimes called “model distillation” — occupies a gray area in intellectual property law that has drawn scrutiny as the AI industry matures. OpenAI’s terms of service generally prohibit using its outputs to develop competing models, though enforcement of such provisions remains legally untested at scale.

The report drew attention given Musk’s repeated public criticism of OpenAI’s practices, including allegations that the organization abandoned its nonprofit mission and operates in ways harmful to the public interest. His lawsuit against OpenAI, filed in early 2024 and later refiled with expanded claims, has positioned him as a critic of the company’s business practices.

For xAI, which has raised billions in venture capital and positioned Grok as a competitor to ChatGPT and Claude, the admission could invite legal challenges from OpenAI or complicate Musk’s public positioning in his ongoing disputes with the company.

The practice of using one AI model’s outputs to train another has become a flashpoint in the industry. Several AI companies have accused competitors of similar conduct, and the legal framework governing such activity remains largely unsettled in U.S. courts.

Neither xAI nor OpenAI immediately responded to requests for comment on the implications of the reported admission.

The development adds another dimension to the legal landscape surrounding AI training data — a subject already under scrutiny from courts, regulators, and lawmakers across the United States.