Murati testifies Altman lied about model safety reviews
openai former cto mira murati testified under oath that ceo sam altman lied to her regarding safety protocols for a new gpt model during the musk v. altman trial on wednesday. murati stated altman falsely claimed the legal department approved skipping the deployment safety board, contradicting general counsel jason kwon. she also criticized altman for undermining her ability to lead and creating a complex management environment, describing his behavior as part of a pattern of lying that other execs like ilya sutskever also alleged.
OpenAI former CTO Mira Murati testified under oath that CEO Sam Altman lied to her regarding safety protocols for a new AI model. During the ongoing Musk v. Altman trial, Murati stated that Altman falsely claimed the legal department approved bypassing the company's deployment safety board for a GPT model. Murati contradicted this in a video deposition, confirming a "misalignment" between Altman's assertions and those of General Counsel Jason Kwon, which led her to ensure the model underwent board review regardless. Murati also described Altman’s management style as detrimental, citing instances where he undermined her authority and pitted executives against one another.