Mira Murati just dropped a bombshell in the Musk v. Altman trial. OpenAI's former CTO testified under oath that Sam Altman straight up lied to her about safety standards for a new AI model.
The specific claim: Altman told Murati that OpenAI's legal department said a new model didn't need to go through the company's deployment safety board. When asked in her video deposition if Altman was telling the truth, Murati's answer was a flat "no."
This isn't just internal drama. It goes directly to the heart of AI safety governance at the company building some of the most powerful AI systems in the world. If leadership is bypassing safety reviews, that's a red flag for anyone relying on these tools.
Murati also said Altman made her work more difficult during her time as CTO. She left OpenAI earlier this year after playing a key role in developing GPT-4 and other major releases.
For context, this testimony came up during Elon Musk's lawsuit against Altman and OpenAI. Musk has argued the company abandoned its original nonprofit mission in favor of profit, and internal conflicts over safety procedures fit right into that narrative.
If you're building products on OpenAI's APIs or making business decisions based on their safety claims, this testimony matters. It suggests there may be gaps between the company's public safety commitments and what actually happens behind closed doors.
