Blow-by-blow account reveals that drawn-out tensions led to the ouster of OpenAI CEO Sam Altman

OpenAI’s board made a surprising move by dismissing Chief Executive Officer Sam Altman just before Thanksgiving. However, the board provided minimal information, only stating that Altman, the head of the AI startup, had not been consistently transparent with the directors.

The statement, devoid of details, was the opening flight in a power struggle that was played out almost entirely behind closed doors. Privately, Altman and the board debated what to say publicly and when, according to people familiar with the situation. At one point in discussions about Altman’s possible return as CEO, he offered to publicly apologize for misrepresenting the views of some board members in discussions as he lobbied for the executive’s ouster, the people said.

But the board was concerned that an apology for one incident could make it seem like it was the only reason he was fired, one person said, and leaders believed the problems ran deeper.

The board has declined to elaborate on its reasoning, citing an ongoing independent investigation, but more details about the decision-making are emerging. According to several people familiar with the board’s thinking, who asked not to be identified to discuss private conversations, the executive move was the culmination of months of deliberation over Altman’s strategic maneuvering and a perceived lack of transparency in his communications with executives.

In a statement to Bloomberg News, an OpenAI spokesperson told Bloomberg News: “We look forward to the results of the board’s independent review. Our primary focus remains developing and publishing useful and safe AI, and supporting the new board as they work to improve our governance structure. On Friday, OpenAI’s chairman said two WilmerHale lawyers would lead to a review.

Board members had started talking about firing Altman earlier in the fall, according to one of the people. At the time, the group included Altman, president Greg Brockman, and four people who eventually ousted him as CEO: OpenAI chief scientist Ilya Sutskever, Quora Inc. CEO Adam D’Angelo, AI academic Helen Toner, and entrepreneur Tasha McCauley. The company has an unusual structure — an unpaid non-profit board oversees the AI startup juggernaut. The main goal of the instructors is to guide the development of artificial intelligence safely.

The board had heard from some OpenAI executives who had problems with Altman, said one person familiar with the executives’ thinking. But employees approached board members cautiously, fearing possible repercussions from Altman when they learned they had spoken out against him, the person said. The Washington Post previously reported some details of the worker unrest.

The company’s spokesperson emphasized in his statement that after Altman’s dismissal, OpenAI’s management team unanimously asked Altman to return as CEO and the board to resign. “The strong support of his team underscores that he is an effective CEO who is open to diverse perspectives, ready to solve complex challenges and demonstrate that he cares about his team,” the statement said.

As the board considered Altman’s leadership, Sutskever’s concerns grew. Before joining OpenAI, the Israeli-Canadian computer scientist worked at Google Brain and was a researcher at Stanford University. In July, he formed a new team at the company to put the “super-smart” artificial intelligence systems of the future under control. And in October, Sutskever’s responsibilities at the company were reduced, reflecting friction between him and Altman and Brockman. Sutskever later appealed to the board and won over some members, including Toner, director of strategy at Georgetown’s Center for Security and Emerging Technologies.

Also in October, Altman tried to unseat Toner. It was a research paper he co-authored that criticized OpenAI’s security practices. After Altman raised concerns about the paper, Toner sent an email to other board members alerting them to the study and offering to answer questions about it, one of the people said. One concern Altman raised, the person said, was that with OpenAI under scrutiny — due to an ongoing FTC investigation — it would look bad for a board member to say anything critical of the company, as regulators might conclude the company had deeper problems. OpenAI.

Altman himself also spoke with some board members. It was these conversations that proved particularly problematic, according to several people. They said that in some conversations with executives, Altman misrepresented the views of others and suggested that other executives agreed with him that Toner should resign after that. Some of these conversations were previously reported in the New Yorker and New York Times.

At one point, one of them said Altman told some directors that McCauley had said, “Helen obviously has to go,” a characterization McCauley disputed. Directors thought those conversations represented Altman’s manipulative behavior, the people said.

An OpenAI spokesperson said this account “differs significantly from Sam’s recollection of these conversations.”

In a memo sent to OpenAI staff the day after Altman’s ouster, Chief Operating Officer Brad Lightcap said Altman’s firing “was due to a breakdown in communications between Altman and the board.” When asked in a recent interview with Bloomberg how he plans to ensure there is no breakdown in communication with board members in the future, Altman said he didn’t want to “establish exactly what happened,” but that “good communication is a very important thing.”

Board members were also concerned that the CEO wasn’t always fully transparent—and if they couldn’t get a clear picture from Altman, they couldn’t effectively monitor him. That, in turn, would make it impossible to do his job under the supervision of one of the world’s most important technology leaders.

When the board decided to fire him, they acted quickly and without notice. This horrified OpenAI’s investors and many of its employees. Since there was no clear explanation of the reason for Altman’s dismissal, almost the entire company’s approximately 770-person workforce quickly signed a letter threatening to resign if the CEO was not reinstated. One thing the board members favoring Altman’s ouster didn’t count on, one person said, was how quickly so many OpenAI employees would rally to Altman’s side.

One of them was Sutskever, who reversed his decision to help fire Altman. During negotiations regarding Altman’s return, Altman demanded a statement from the board clearing him of wrongdoing, people familiar with the matter said. Directors were unwilling to give in to this and other demands, Bloomberg reported. But within days, Altman was reinstated.

Since Altman’s ouster and return, both Toner and McCauley have stepped down. The only remaining member of the volunteer board that existed before November 17th is Quora’s D’Angelo. Outgoing executives pushed to keep him, one person said, in part because they wanted someone at the company who remembered what happened during the company’s chaotic leadership battle and the events leading up to it.