Mira Murati expresses distrust in Sam Altman’s statements during court proceedings.

Mira Murati, the former CTO of OpenAI, has testified that CEO Sam Altman misled her regarding the safety protocols for a new AI model. In a video deposition presented during the ongoing Musk v. Altman trial, Murati stated that Altman incorrectly claimed that the legal team had decided the new AI model didn’t require review by the company’s deployment safety board. When asked if Altman was truthful in that assertion, Murati replied, “No.”

She expressed that her experience at OpenAI was challenging due to Altman’s management style. “My criticism is entirely about management,” she noted. “I had an incredibly hard job to do in a very complex organization. I was asking Sam to lead with clarity and not undermine my ability to perform my role.”

One specific instance involved the safety protocols for one of OpenAI’s GPT models. After her conversation with Altman, Murati consulted Jason Kwon, who became OpenAI’s general counsel in 2021 and is now the chief strategy officer. She identified a “misalignment” between Kwon’s and Altman’s statements, confirming that their messages did not align. To ensure safety, she decided to have the model reviewed by the board.

This isn’t the first time Altman has faced accusations of dishonesty. Cofounder Ilya Sutskever, in a memo to OpenAI’s board, described Altman as having a “consistent pattern of lying, undermining his execs, and pitting them against one another.” Former board member Helen Toner also mentioned in a podcast that executives had presented evidence of Altman being “manipulative” in various situations. Murati concurred with these observations, noting that Altman often created divisions among executives and undermined her efforts.

When the board dismissed Altman, they stated he “was not consistently candid in his communications with the board, hindering its ability to fulfill its responsibilities. The board no longer has confidence in his ability to lead OpenAI.” In response to a question about the board’s loss of trust, Altman remarked, “That will be a better question for them.”

Following Altman’s dismissal, Murati briefly served as interim CEO. However, she criticized the board’s decision, stating that “OpenAI was at catastrophic risk of falling apart.” She left the company in 2024 and subsequently launched her own competitor, Thinking Machines Lab.