Altman testifies about ‘hair-raising’ AI safety chat with Musk

Altman’s testimony caps off years of animosity between two tech billionaires who bonded over a shared vision for AI development before falling out over the company’s direction. Photo by John MACDOUGALL and SAUL LOEB/AFP via Getty Images

Article content

OpenAI’s Sam Altman testified that he was “extremely uncomfortable” with Elon Musk’s insistence that he have complete control over a proposed for-profit subsidiary of OpenAI back in 2017.

Article content

Altman told jurors that Musk said he only wanted control of the entity in its early days, but that the billionaire was unwilling to sign a contract to that effect, frustrating fellow co-founders of the artificial intelligence startup.

Article content

Article content

Article content

“A particularly hair-raising moment was when my co-founders asked Mr. Musk, ‘If you have control, what happens when you die?’” Altman said. “He said something like, ‘I haven’t thought about it a ton, but maybe I should pass it to my children.’”

Article content

Article content

“I didn’t feel comfortable with that,” Altman added.

Article content

The OpenAI chief executive took the witness stand Tuesday morning in the final days of the high-profile trial. Musk has accused Altman and company president Greg Brockman of seeking to enrich themselves by abandoning the nonprofit’s altruistic mission and converting it into a for-profit business company with billions of dollars in investments from Microsoft Corp.

Article content

Musk previously testified that he started losing confidence in Altman and Brockman during the 2017 negotiations about OpenAI’s future. He was OpenAI’s biggest financial backer at the time, and thought his equity in its for-profit venture should reflect that.

Article content

Musk testified that he gave OpenAI roughly US$38 million in the form of quarterly donations and rent payments for its office space — far less than the US$1 billion he initially pledged.

Article content

Article content

Altman told jurors on Tuesday that he contributed US$3.75 million to the nonprofit when it first started, and then gave “a multiple of that” in equity interests he held to employees in its early days. Altman has repeatedly said he doesn’t currently have a direct equity stake in OpenAI.

Article content

Article content

Altman said he and the other co-founders also devoted significant time and effort toward launching OpenAI, and he felt frustrated that those efforts were not reflected in Musk’s proposals for the equity split in the for-profit venture.

Article content

“It was totally true that Elon was proposing to put in a lot of money, and Elon had you know a lot of brand value,” Altman said. “But I really wanted to stand up for Greg and Ilya,” referring to OpenAI co-founder Ilya Sutskever, who went on to become its chief scientist for several years.

Article content

“One of the most special things about OpenAI has been we have incredible team spirit and teamwork,” Altman told jurors. “And I remember in these conversations, I felt that Elon just really deeply didn’t get that and wasn’t valuing it.”

Article content

Musk departed OpenAI’s board in 2018, and stopped paying rent on its office space in 2020. Altman said Musk’s departure affected OpenAI “up and down” and raised questions about its fundraising capabilities without him on board.