What happened at OpenAI over the past five days could be described in many ways: A juicy boardroom drama, a tug of war over one of America’s biggest start-ups, a clash between those who want A.I. to progress faster and those who want to slow it down.
But it was, most importantly, a fight between two dueling visions of artificial intelligence.
In one vision, A.I. is a transformative new tool, the latest in a line of world-changing innovations that includes the steam engine, electricity and the personal computer, and that, if put to the right uses, could usher in a new era of prosperity and make gobs of money for the businesses that harness its potential.
In another vision, A.I. is something closer to an alien life form — a leviathan being summoned from the mathematical depths of neural networks — that must be restrained and deployed with extreme caution in order to prevent it from taking over and killing us all.
With the return of Sam Altman on Tuesday to OpenAI, the company whose board fired him as chief executive last Friday, the battle between these two views appears to be over.
Team Capitalism won. Team Leviathan lost.
OpenAI’s new board will consist of three people, at least initially: Adam D’Angelo, the chief executive of Quora (and the only holdover from the old board); Bret Taylor, a former executive at Facebook and Salesforce; and Lawrence H. Summers, the former Treasury secretary. The board is expected to grow from there.
OpenAI’s largest investor, Microsoft, is also expected to have a larger voice in OpenAI’s governance going forward. That may include a board seat.
Gone from the board are three of the members who pushed for Mr. Altman’s ouster: Ilya Sutskever, OpenAI’s chief scientist (who has since recanted his decision); Helen Toner, a director of strategy at Georgetown University’s Center for Security and Emerging Technology; and Tasha McCauley, an entrepreneur and researcher at the RAND Corporation.
Mr. Sutskever, Ms. Toner and Ms. McCauley are representative of the kinds of people who were heavily involved in thinking about A.I. a decade ago — an eclectic mix of academics, Silicon Valley futurists and computer scientists. They viewed the technology with a mix of fear and awe, and worried about theoretical future events like the “singularity,” a point at which A.I. would outstrip our ability to contain it. Many were affiliated with philosophical groups like the Effective Altruists, a movement that uses data and rationality to make moral decisions, and were persuaded to work in A.I. out of a desire to minimize the technology’s destructive effects.
This was the vibe around A.I. in 2015, when OpenAI was formed as a nonprofit, and it helps explain why the organization kept its convoluted governance structure — which gave the nonprofit board the ability to control the company’s operations and replace its leadership — even after it started a for-profit arm in 2019. At the time, protecting A.I. from the forces of capitalism was viewed by many in the industry as a top priority, one that needed to be enshrined in corporate bylaws and charter documents.
But a lot has changed since 2019. Powerful A.I. is no longer just a thought experiment — it exists inside real products, like ChatGPT, that are used by millions of people every day. The world’s biggest tech companies are racing to build even more powerful systems. And billions of dollars are being spent to build and deploy A.I. inside businesses, with the hope of reducing labor costs and increasing productivity.
The new board members are the kinds of business leaders you’d expect to oversee such a project. Mr. Taylor, the new board chair, is a seasoned Silicon Valley deal maker who led the sale of Twitter to Elon Musk last year, when he was the chair of Twitter’s board. And Mr. Summers is the Ur-capitalist — a prominent economist who has said that he believes technological change is “net good” for society.
There may still be voices of caution on the reconstituted OpenAI board, or figures from the A.I. safety movement. But they won’t have veto power, or the ability to effectively shut down the company in an instant, the way the old board did. And their preferences will be balanced alongside others’, such as those of the company’s executives and investors.
That’s a good thing if you’re Microsoft, or any of the thousands of other businesses that rely on OpenAI’s technology. More traditional governance means less risk of a sudden explosion, or a change that would force you to switch A.I. providers in a hurry.
And perhaps what happened at OpenAI — a triumph of corporate interests over worries about the future — was inevitable, given A.I.’s increasing importance. A technology potentially capable of ushering in a Fourth Industrial Revolution was unlikely to be governed over the long term by those who wanted to slow it down — not when so much money was at stake.
There are still a few traces of the old attitudes in the A.I. industry. Anthropic, a rival company started by a group of former OpenAI employees, has set itself up as a public benefit corporation, a legal structure that is meant to insulate it from market pressures. And an active open-source A.I. movement has advocated that A.I. remain free of corporate control.
But these are best viewed as the last vestiges of the old era of A.I., in which the people building A.I. regarded the technology with both wonder and terror, and sought to restrain its power through organizational governance.
Now, the utopians are in the driver’s seat. Full speed ahead.
