## Former OpenAI CTO Alleges Sam Altman Lied Under Oath in Musk v. Altman Trial
Mira Murati, OpenAI's former chief technology officer, has testified under oath that CEO Sam Altman lied to her about the safety protocols governing a new AI model, according to video deposition footage presented during the ongoing Musk v. Altman trial. Murati stated that Altman falsely claimed OpenAI's legal department had determined the model did not require review by the company's deployment safety board. When asked whether Altman was telling the truth in that statement, Murati's response was direct: "No."

The testimony, shown to the court on Wednesday, marks one of the most significant insider accounts of internal decision-making at OpenAI. Murati, who served as CTO before departing the company, told investigators that Altman's misrepresentations complicated her ability to manage safety processes during her tenure. The specific incident centers on whether a new AI model required evaluation by OpenAI's formal deployment safety board—a gatekeeping mechanism designed to catch potential harms before release. Murati's account suggests that determination was misrepresented to bypass that review.

The deposition arrives amid heightened legal scrutiny of OpenAI's leadership and governance structure. Elon Musk, a co-founder who later departed and publicly clashed with Altman, has brought claims against the company that touch on its original nonprofit mission and allegations of operational misconduct. Murati's testimony adds a named, sworn account to those allegations, drawing direct lines between Altman's communications and what she perceived as false representations about internal safety processes. The trial continues to examine whether OpenAI's leadership consistently prioritized speed of deployment over established safety guardrails.
---
- **Source**: The Verge
- **Sector**: The Lab
- **Tags**: openai, sam altman, mira murati, ai safety, musk v altman
- **Credibility**: unverified
- **Published**: 2026-05-06 19:31:39
- **ID**: 79874
- **URL**: https://whisperx.ai/en/intel/79874