Sam Altman's shock return as chief executive of OpenAI late on November 22 – days after being sacked – caps a chaotic period that highlighted deep tensions at the heart of the artificial intelligence community.KIRILL KUDRYAVTSEV/AFP/Getty Images
Gus Carlson is a U.S.-based columnist for The Globe and Mail.
When Charlie Chaplin, Mary Pickford, Douglas Fairbanks and D.W. Griffith founded film studio United Artists in 1919, rival Metro Pictures head Richard A. Rowland famously said that “the lunatics have taken over the asylum.”
It’s not that they weren’t talented actors and recognized stars – they were, and Mr. Rowland acknowledged that. It’s just that they weren’t qualified or experienced enough to oversee the complicated and competitive business of running a film studio.
Fast-forward more than a century to the OpenAI boardroom, where Mr. Rowland’s observation is as relevant now as it was then. And, remarkably, the governance storyline that unfolded last week at the startup responsible for ChatGPT is so farcical it could have been scripted by Mr. Chaplin, the comedic genius, himself.
The saga of the board’s sudden dismissal of chief executive officer Sam Altman – and his phoenix-like reinstatement to the job on Tuesday – has dominated tech headlines for several days. But what has really captivated observers is the breathtakingly bizarre behaviour of OpenAI’s directors, whipsawing in their decision-making, sending mixed and conflicting messages to the market, contradicting each other and themselves and at every turn finding ways to make a bad situation worse.
As stunning as the Keystone Kops-like boardroom antics have been, they should come as no surprise. Seen through Mr. Rowland’s prism, OpenAI’s directors may be competent in their respective areas of expertise – technology, venture capital, entrepreneurship, academics – but they were clearly way out of their depth when it comes to governance.
To the relief of many, in announcing Mr. Altman’s return, OpenAI also unveiled a new initial board that includes former Salesforce co-CEO Bret Taylor as chair, former U.S. treasury secretary Larry Summers and Adam D’Angelo as directors. Mr. D’Angelo was part of the original board that fired Mr. Altman.
The OpenAI experience is an all-too-familiar cautionary tale illustrating that properly credentialed and experienced directors are critical to an effective and mature governance structure. Proper governance requires grown-ups in the boardroom, and too often that is not the case.
The shortfall that rocked OpenAI is not insignificant or as narrow as the fate of a single film studio. OpenAI and Mr. Altman have been considered beacons – and, in some ways, the guardians of humanity – in the fast-emerging world of artificial intelligence, which many fear is already a runaway train.
The board’s handling of the situation did not instill confidence that the organization could guard itself, let alone humanity. They proved once again the time-honoured truth that most corporate crises are self-inflicted.
Even the messaging around Mr. Altman’s dismissal was fumbled. First, the suggestion was that he wanted to monetize AI more quickly than the board was comfortable doing; then it was his lack of transparency and poor communications that prompted his ouster.
Then, almost in the same breath, the board suddenly promoted OpenAI’s chief technology officer, Mira Murati, as his replacement. Just hours later, Ms. Murati signed a letter along with hundreds of her colleagues calling for Mr. Altman’s return and the resignation of the board – then threatened to quit.
The head-scratching was far from done. Director Ilya Sutskever, an OpenAI co-founder and member of the board who reportedly pushed for Mr. Altman’s dismissal, surprised many by signing the letter as well and publicly apologized for his role in the convoluted tale.
The back-and-forth had some observers anticipating the board members might step down in a fit of self-awareness, having realized that they were the cause of the crisis, but that has yet to happen despite further noise from OpenAI employees who threatened to walk if things weren’t put back in order.
Amid the turmoil, Microsoft, which has a strong partnership with OpenAI and was blindsided by Mr. Altman’s dismissal, said it would hire Mr. Altman and former OpenAI president Greg Brockman to lead an in-house AI research team.
The return of Mr. Altman to lead OpenAI and the appointment of a new board may end the week of frenzy at the startup, but they will not immediately answer the big questions: How much damage has been done by the board’s conduct to OpenAI’s position as a leader in the AI industry? And how will it affect the relationship with Microsoft? Microsoft has already said it will demand a board seat to avoid further surprises.
If you are at all worried about the influence artificial intelligence will have on our lives, the OpenAI fiasco will not help you sleep at night. Even with Mr. Altman’s return and a reconstituted board, there is a lot of work to be done to restore confidence that the organization can be trusted with something as critically important as helping manage the evolution of a powerful technology that will change the way we live and work.