Sam Altman just had the worst week of testimony in the Musk v. OpenAI trial. And it was not Elon Musk who did the damage.

It was his own people.

Over the last two days of the Oakland federal trial, four former OpenAI leaders took the stand or had their depositions played for jurors. Former CTO Mira Murati. Former safety researcher Rosie Campbell. Former board members Helen Toner and Tasha McCauley. Each testified independently. And each painted essentially the same picture: Sam Altman created chaos, told different people different things, bypassed safety processes, and ran OpenAI in a way that made it nearly impossible to trust what he said.

One person saying this under oath is gossip. Four people saying it under oath, independently, is a pattern.

Murati: "He Was Creating Chaos"

Murati's video deposition was the centerpiece. She testified that Altman told her OpenAI's legal department said it was not necessary for the safety board to review a new model before release. She checked with Jason Kwon, then head of legal (now chief strategy officer), who contradicted Altman's claim. "I confirmed that what Jason was saying and what Sam was saying were not the same thing," Murati told the court.

Her broader testimony was even more damaging. "My concern was about Sam saying one thing to one person and completely the opposite to another person," she said, per Reuters. She described Altman as "creating chaos" and called OpenAI during the 2023 board crisis "at catastrophic risk of falling apart." According to Forbes, she testified that even after his reinstatement, Altman continued creating a "very difficult and chaotic environment" through inconsistent messaging and decision paralysis on big issues.

Campbell: "No Home" for Safety

Rosie Campbell joined OpenAI's AGI readiness team in 2021 and left in 2024 after her team was dissolved. She testified that she believed OpenAI was abandoning its commitment to safety during her time there. According to Business Insider, Campbell told jurors that she "felt she had no home" at the company. She described a shift from a research and safety-focused organization into a product-focused one, where launches took priority over the careful evaluation that had once defined the company's identity.

Campbell's testimony cuts differently from Murati's. Murati's complaint is about management chaos and dishonesty. Campbell's complaint is about mission drift. OpenAI was not founded to ship products on quarterly cycles. It was founded to develop safe artificial general intelligence. Campbell is saying, under oath, that those two things became incompatible.

Toner and McCauley: "A Culture of Lying"

Former board members Helen Toner and Tasha McCauley went further. McCauley testified that she voted to fire Altman after months of complaints from senior employees about what she called a "culture of lying and a culture of deceit." She said Altman wanted to "control the board" and resisted oversight. Toner, who also voted to fire Altman, testified that he "would put words in other people's mouths" and claim false consensus during board discussions.

Both described a board that was systematically kept in the dark. Shivon Zilis, another former board member and mother of Musk's children, separately testified that the entire board was concerned about the lack of communication around the ChatGPT launch, and flagged potential conflicts of interest in a deal between OpenAI and Helion Energy, where both Altman and Brockman had personal investments.

Why This Week Matters

Altman himself admitted in a recent blog post that he is "not proud of being conflict-averse" and acknowledged the trait caused "great pain" for the company. That concession exists in writing. Now it exists alongside sworn testimony from four separate leaders who lived through the consequences.

The legal question in this trial is narrow: did OpenAI breach agreements with Musk? But the testimony this week has painted something much broader. It is not just about Musk's $150 billion lawsuit anymore. It is about whether the person running the company that may shape the future of artificial intelligence can be trusted to tell the truth to the people working alongside him.

Murati said chaos. Campbell said mission drift. McCauley said a culture of lying. Toner said fabricated consensus. Four people, four angles, one conclusion.

Altman is expected to take the stand in the coming days. He will need more than conflict aversion to answer what this week put on the record.

Sources: Gizmodo, Reuters, Forbes, Business Insider, The Verge, NBC Bay Area, India Today, Livemint