A former chief technology officer just told a court under oath something that Senate investigators have spent months trying to confirm: Sam Altman lied about safety protocols at OpenAI.
Mira Murati, who served as OpenAI's CTO before her departure, delivered sworn testimony during the Musk v. Altman trial that directly contradicts Altman's public account of internal safety governance. When asked whether Altman told her the truth when he claimed OpenAI's legal department had cleared a new AI model from required safety board review, Murati answered with a single word: "No."
This is not anonymous sourcing. This is not industry speculation. This is a senior executive with direct knowledge testifying under penalty of perjury that the CEO of the world's most influential AI company made a false statement about safety procedures.
The implications are stark. For months, Senate investigators probing OpenAI's internal governance have faced a wall of denial. Altman's representatives have consistently rejected claims that safety protocols were circumvented. Board members who raised concerns were marginalized or removed. Now the person who oversaw technical development at the company has provided the first direct, named corroboration of what critics have alleged.
Murati's testimony extends beyond the specific incident. She stated that during her tenure, Altman made her work more difficult—a striking admission from someone who held the most technically demanding role at the organization. The combination of testimony paints a picture of systematic pressure on safety personnel, not isolated miscommunication.
The Senate Commerce Committee has been building a case around AI safety governance for over a year. They have interviewed former employees, reviewed internal documents, and issued subpoenas for communications. What they lacked was direct corroboration from someone with executive-level knowledge who could testify toAltman's personal involvement in misrepresenting safety procedures.
Murati just provided that.
Committee members should act immediately. Murati's deposition is a matter of public record in an ongoing civil trial, but sworn testimony before Congress carries different legal weight and would compel fuller disclosure. The question is no longer whether there is credible evidence to pursue this inquiry. The question is whether Senate investigators will exercise their subpoena power while the trail is fresh and witnesses are still willing to speak.
Altman has built his public persona around being the responsible face of AI development. Murati's testimony suggests that image may have been constructed on a foundation of misrepresentations to his own board and executives. The Senate owes the public answers—and Murati just gave them the opening to get them.