Sam Altman returns to OpenAI board after external review
OpenAI is bringing Sam Altman back to its board after WilmerHale’s review of his November ouster. The investigation found product safety was not the main cause of the crisis.
Sam Altman is returning to OpenAI’s board of directors after an external investigation into his abrupt ouster in November concluded that his conduct did not warrant removing him from the company. The review, conducted by the law firm WilmerHale, also found that the safety of OpenAI’s products was not a primary reason for the decision.
The announcement formally closes a corporate governance crisis that exposed a difficult tension at the company behind ChatGPT: OpenAI wants to develop increasingly powerful artificial intelligence systems, but its structure was created with a mission to ensure that the technology benefits humanity.
An investigation into the November episode
The previous board removed Altman on November 17, 2023, saying he had not always been candid in his communications with directors. The decision triggered five days of chaos: nearly the entire workforce threatened to leave OpenAI, Microsoft offered to hire Altman and other executives, and Altman returned as chief executive on November 22.
WilmerHale interviewed dozens of people and reviewed more than 30,000 documents to reconstruct what happened, according to OpenAI. Its conclusion was that Altman’s removal was not primarily driven by concerns about model safety, the pace of development, the company’s finances, or its communications with investors, customers or partners.
The report attributed the conflict to the breakdown of the relationship and trust between the former board and Altman. It also found that the directors acted in good faith and within the scope of their authority, although they failed to anticipate the impact the decision would have on the company’s stability.
That distinction matters. From the outside, the crisis was seen as a potential clash between those who favored moving quickly on generative AI and those calling for greater caution. The investigation does not resolve the debate over the risks posed by increasingly capable models, but it does rule out those concerns as the central, substantiated reason for Altman’s removal.
A revamped board with closer ties to business
Alongside Altman’s return, OpenAI is adding Bret Taylor, former co-CEO of Salesforce; Sue Desmond-Hellmann, former CEO of the Bill & Melinda Gates Foundation; and Nicole Seligman, former executive vice president of Sony, to the board. Adam D’Angelo, the founder of Quora, and Larry Summers, a former U.S. Treasury secretary, will remain.
The new lineup brings together experience in major technology companies, corporate law, public policy and nonprofit organizations. Taylor will chair the board, a particularly important role after the departure of many of the members who made the November decision.
OpenAI retains an unusual structure: its nonprofit entity controls the commercial company that sells access to its models. In theory, that arrangement allows the public mission to take priority over financial interests. In practice, the crisis showed that a board with the power to remove the chief executive also needs clearer procedures for managing internal disagreements with far-reaching consequences.
More safeguards, but a fundamental question remains open
The company has announced governance changes, including strengthening its conflict-of-interest policy and creating a confidential reporting line for employees and others to raise concerns. It also plans to expand the board in the future.
The measures are intended to prevent a dispute between management and directors from escalating again without mediation mechanisms or adequate planning. But they do not, on their own, resolve the question that has followed OpenAI since its founding: who decides, and according to what criteria, when an AI product is ready for large-scale deployment.
For users and business customers, the outcome provides a signal of continuity at a time when OpenAI is competing with Google, Anthropic, Meta and other players to lead generative AI. For the industry, it leaves a less comfortable lesson: AI safety does not depend solely on technical testing. It also requires boards capable of overseeing executives without paralyzing an organization that has become infrastructure for millions of people and businesses.
