The end of an investigation into the tumultuous firing of Sam Altman from OpenAI more than three months ago represents a resounding victory for the high-profile chief executive as he moves to reassert control of the artificial intelligence company he helped create.
OpenAI, in a news conference on Friday, said that Mr. Altman, who returned to OpenAI just five days after he was ousted in November, didn’t do anything to justify his departure and reclaim a role at the company that still hasn’t escaped. he: a seat on the company’s board of directors.
The expulsion of Mr. Altman shocked Silicon Valley and damaged the future of one of the tech industry’s most influential start-ups. It also asked whether OpenAI – with or without Mr. Altman in charge – is poised to carry the banner for the tech industry’s fervent focus on artificial intelligence.
When he returned to OpenAI in November, Mr. Altman did not regain his board seat while agreeing to an investigation into his conduct and the board’s actions. Two members who voted for his removal agreed to step down; their replacements, from outside the company, managed the investigation by the law firm, WilmerHale. Bret Taylor, chairman of OpenAI’s board, said at the news conference that the long-awaited report about the episode was finished, but the company did not release the report.
The company said the law firm’s report found that OpenAI’s board acted within its broad discretion to terminate Mr. Altman, but also found that his behavior did not warrant removal.
“The special committee recommended and the entire board expressed their full confidence in Mr. Altman and Mr. Brockman,” Mr. Taylor said, referring to Greg Brockman, the company’s president who quit in protest after being fired Mr. Altman. “We are excited and united in our support for Sam and Greg.”
OpenAI also moved to address concerns about the lack of diversity on the board by adding three women as directors: Sue Desmond-Hellmann, the former chief executive of the Bill & Melinda Gates Foundation; Nicole Seligman, Sony’s former general counsel; and Fidji Simo, the chief executive of Instacart.
Mr. Taylor, who was one of those named to OpenAI’s board in November, said the board will continue to expand.
With the report and the additions to the board, OpenAI’s leadership hopes to move past the controversy of Mr. Altman’s ouster. The incident raised many questions about his leadership and the unusual structure of the San Francisco company — a nonprofit board overseeing a for-profit company.
But because it did not release the report, OpenAI left many questions unanswered about the company. Some insiders questioned whether Mr. Altman had too much control over how the investigation was handled.
“As we told the investigators, deception, manipulation, and resistance to scrutiny should be unacceptable,” Helen Toner and Tasha McCauley, the two OpenAI board members who left late last year, said. said in a statement. “We hope the new board will do its job in managing OpenAI and holding it accountable to the mission.”
Mr. Taylor appeared next to Mr. Altman at the news conference on Friday. After announcing the new board members, he said the review found that the previous board acted in good faith in removing Mr. Altman but did not anticipate the challenges that would arise from his dismissal.
“The review determined that the board’s decision did not stem from concern about the safety or security of the product,” Mr. Taylor said. “This is simply a breakdown of trust between the board and Mr. Altman.”
After Mr. Taylor completed his prepared remarks, Mr. Altman praised the resilience of the company and its partners during and after his ouster. “I’m glad it’s all over,” he said.
OpenAI provided a six-paragraph summary of the report. It said WilmerHale reviewed 30,000 documents and conducted dozens of interviews, including with former OpenAI board members.
It was found that the previous board was accurate in its rationale and public explanation for the expulsion of Mr. Altman for not being “consistently honest in his communications with the board.” It also said the board did not expect its action to destroy the company.
The company said WilmerHale gave oral briefings on the report, which will not be made public, to Mr. Taylor and Lawrence H. Summers, the former Treasury secretary who were also added to the board in November.
Mr. Taylor said OpenAI has made several changes intended to improve the way the company operates, including new governance guidelines for the board, a new conflict of interest policy and a whistle-blower hotline.
OpenAI’s summary of the report did not provide insight into the concerns that the company’s senior leaders brought to the previous board about Mr. Altman. Before his dismissal, Ilya Sutskever, OpenAI’s chief scientist, and Mira Murati, OpenAI’s chief technology officer, expressed concerns about Mr. Altman’s management style, including what they characterized as his history of manipulative behavior, The New York Times reported.
Dr. Sutskever, through a lawyer, called those claims “false.” said Ms. Murati in a company Slack post on Thursday he shared the same feedback with the board that he gave directly to Mr. Altman, but said he never contacted the board to share those concerns.
“I’m happy that the independent review is over and we can all move forward together,” said Ms. Murati on Friday in a post on X, formerly known as Twitter.
OpenAI is still being investigated by the Securities and Exchange Commission over the board’s actions and the possibility that Mr. Altman misled investors. Companies that hire outside law firms often turn the report over to public investigators after completion. A spokesman for OpenAI’s board declined to say whether it would provide the report to the SEC
(The New York Times sued OpenAI and Microsoft in December for copyright infringement of news content related to AI systems.)
OpenAI, valued at more than $80 billion in its most recent financing round, is a leader in generative AI, technologies that can generate text, images and sounds. Many believe that generative AI could change the technology industry as radically as the web browser did about three decades ago. Others worry that the technology could cause serious harm, helping to spread disinformation online, displacing countless jobs and perhaps even threatening the future of humanity.
After OpenAI released the online chatbot ChatGPT in late 2022, Mr. Altman A year later, the board unexpectedly removed him, saying it no longer had confidence in his ability to run the company.
The board has shrunk to six people: three founders and three independent members. Along with three outsiders, Dr. Sutskever, one of the founders of OpenAI, voted to fire Mr. Altman as chief executive and chairman of the board, saying he has not been “consistently honest in his communications.”
Mr. Brockman, another founder, resigned from the company in protest. Days later, Dr. Sutskever regretted his decision to remove Mr. Altman and effectively stepped down, leaving three independent members standing in opposition to Mr. Altman.
OpenAI was founded as a nonprofit in 2015, before Mr. Altman created a for-profit subsidiary three years later and raised $1 billion from Microsoft. The nonprofit’s board, whose stated mission is to develop AI for the benefit of humanity, retains complete control of the new subsidiary. Investors, including Microsoft, had no legal say in who ran the company.
In an effort to resolve the chaos and bring back Mr. Altman in the company, he and the board agreed to replace two members of Mr. Taylor, who is a former Salesforce executive. But Mr. Altman was not reinstated to the board. Mr. Taylor and Mr. Summers were charged with overseeing the investigation into Mr. Altman and his firing.
Microsoft, a close partner of OpenAI, has a board observer position, filled by Dee Templeton, the company’s vice president, technology and research partnerships. Microsoft on Friday declined to comment on the board and report.
The new board has faced criticism from corporate governance experts for its lack of diversity. Mr. Taylor told The Times in November that he would fill out the board by adding “qualified, diverse candidates” who embody “the totality of what this mission represents, spanning technology, policy AI safety.”
Karen Weise contributed reporting.
