Even with Sam Altman’s return, OpenAI must prove itself

For now, the OpenAI power struggle that captivated the tech world after co-founder Sam Altman was fired has ended. But how do you interpret?

It seems like OpenAI died and a new, but not necessarily better, startup replaced it. Ex-Y Combinator President Altman returns, but is he justified? OpenAI’s founding philanthropic goals may be co-opted by capitalist interests, as its new board is entirely white and male.

By no means was the old OpenAI perfect.

OpenAI had six board members on Friday morning: Altman, chief scientist Ilya Sutskever, president Greg Brockman, tech entrepreneur Tasha McCauley, Quora CEO Adam D’Angelo, and Georgetown Center for Security and Emerging Technologies director Helen Toner. The board was technically tied to a nonprofit that owned a majority stake in OpenAI’s for-profit side and exercised complete control over its activities, investments, and direction.

OpenAI’s co-founders, including Altman, created its unusual structure with good intentions. The nonprofit’s tiny (500-word) charter requires the board to make decisions to ensure “that artificial general intelligence benefits all humanity,” leaving members to interpret that. Toner reportedly told Altman’s executive team that triggering OpenAI’s collapse “would actually be consistent with the [nonprofit] mission.”This North Star document doesn’t mention “profit” or “revenue.”

At OpenAI, the arrangement worked well for years, so maybe it worked in another universe. Once investors and powerful partners joined, things got more complicated.

Altman’s firing unites Microsoft and OpenAI staff

After the board abruptly fired Altman on Friday without notifying anyone, including most of OpenAI’s 770 employees, backers voiced their dissatisfaction in private and public.

The departure of Altman reportedly “furious” Microsoft CEO Satya Nadella, a major OpenAI collaborator. Another OpenAI backer, Vinod Khosla, said on X (formerly Twitter) that the fund wanted Altman back. Thrive Capital, Khosla Ventures, Tiger Global Management, and Sequoia Capital were considering legal action against the board if Altman’s reinstatement negotiations over the weekend failed.

It appears OpenAI employees were aligned with these investors. Nearly all of them—including Sutskever, who appeared to change his mind—signed a letter threatening mass resignation if the board did not reverse course. Despite job offers from Microsoft and Salesforce, these OpenAI employees had a lot to lose if OpenAI failed.

OpenAI, led by Thrive, was considering selling employee shares to raise its valuation from $29 billion to $80 billion to $90 billion. Altman’s abrupt departure and OpenAI’s questionable interim CEOs made Thrive wary, jeopardizing the sale.

At what cost did Altman win the five-day battle?

After several breathless, hair-pulling days, a resolution has been reached. Altman, along with Brockman, who resigned on Friday in protest of the board’s decision, is back, pending a background investigation into his removal. A new transitional board at OpenAI meets Altman’s request. With investors’ profits capped and the board free to make non-revenue-driven decisions, OpenAI will reportedly remain structured.

Salesforce CEO Marc Benioff said “the good guys” won on X. But that may be premature.

Altman “won,” defeating a board that accused him of “not [being] consistently candid” and possibly prioritizing growth over mission. Altman was alleged to have tried to push Toner off the board after she co-authored a paper criticizing OpenAI’s safety practices. Also, Altman “infuriated” Sutskever by rushing AI-powered features at OpenAI’s first developer conference.

After multiple chances, the board didn’t explain, citing legal issues. The Altman dismissal was unnecessary and histrionic. But depending on how they interpreted their humanistic directive, the directors may have had good reasons for firing Altman.

This directive may be interpreted differently by the new board.

Larry Summers, an economist and former Harvard president, and former Salesforce co-CEO Bret Taylor make up OpenAI’s board. Taylor co-founded many companies, including FriendFeed (acquired by Facebook) and Quip (which Salesforce acquired via acquisition). Summers has deep business and government connections, which OpenAI may benefit from as AI regulation intensifies.

However, if diverse viewpoints were the goal, the directors are not a “win” to this reporter. Although six seats remain unfilled, the initial four set a homogenous tone. In Europe, companies must reserve at least 40% of their board seats for women.

Why some AI experts worry about OpenAI’s new board

I’m not alone in disappointment. Earlier today, AI academics vented on X.

Noah Giansiracusa, a Bentley University math professor and author of a book on social media recommendation algorithms, criticizes the board’s all-male makeup and Summers’ nomination, who has made derogatory remarks about women.

“Whatever one makes of these incidents, the optics are not good, to say the least — particularly for a company that has been leading the way on AI development and reshaping the world we live in,” Giansiracusa texted. “It’s troubling that OpenAI’s main goal is to create AI that ‘benefits all of humanity.’ Recent events make me doubt this since half of humanity is female. Toner most directly represents the safety side of AI, which has been the role of women throughout history, especially in tech: protecting society from great harms while men innovate and rule the world.

Sanford’s AI Lab director Christopher Manning is more charitable than Giansiracusa but agrees:

“The newly formed OpenAI board is presumably still incomplete,” he told. “However, the current board membership, with only white males and no deep knowledge of responsible AI use in human society, is not a promising start for such an important and influential AI company.”

From annotators who label data used to train generative AI models to harmful biases in those trained models, including OpenAI’s, the AI industry is unequal. Fair enough, Summers has worried about AI’s potential impact on livelihoods. Critics I spoke with doubt that OpenAI’s current board will prioritize these issues as much as a more diverse board would.

Why didn’t OpenAI hire an AI ethicist like Timnit Gebru or Margaret Mitchell for the initial board? Were they unavailable? Did they decline? Or did OpenAI never try? We may never know.

Laurene Powell Jobs and Marissa Mayer were considered for OpenAI board seats but rejected due to their closeness to Altman. Condoleezza Rice was also considered but rejected.

OpenAI can prove itself wiser and worldlier by choosing the five remaining board seats—or three if Altman and a Microsoft executive each take one (rumored). Without a more diverse approach, Daniel Colson, director of the AI Policy Institute, may be right: a few people or a single lab can’t be trusted to responsibly develop AI.