Ever since he had a fireside discussion about artificial intelligence with Google co-founder Larry Page more than a decade ago, Elon Musk has had one big fear: that artificial intelligence might eventually destroy humanity.
That was one of the reasons, he often said, that he founded the non-profit AI lab OpenAI with Sam Altman, Greg Brockman and a group of AI researchers. They were going to safely build technology to benefit humanity and protect the world from people like Mr. Page who didn’t believe AI was a threat.
But the nine jurors deciding Mr Musk’s landmark lawsuit against OpenAI are unlikely to hear much about his nightmares. Before he returned to the witness stand for a third day on Thursday, Judge Yvonne Gonzalez Rogers, who is presiding over the trial, told Mr. Musk’s lawyer that she did not want talk of AI’s existential threat to humanity to enter the trial.
“We’re not going to get into the problems of disaster and extinction,” Judge Gonzalez Rogers said. When Mr. Musk’s lead attorney, Steven Molo, began arguing with OpenAI’s lawyer over the matter, the judge raised her voice and insisted they stop bickering.
“I suspect that there are a number of people who do not want to put the future of humanity in Mr. Musk’s hands,” the judge said. “But we’re not going to get into that. We’re just not going to blow it all up for the world to see.”
Crucial to Mr. Musk’s case is whether lawyers can discuss “human extinction.” His lawyers have gone to great lengths to emphasize the existential nature of his concerns, trying to underline that he is trying to protect the world from what OpenAI might create, not just hurt a competitor of his own AI startup. Judge Gonzalez Rogers’ decision to remove that line of questioning could be a blow to Mr. Musk.
In a trial expected to last a month in federal court in Oakland, Calif., he accused OpenAI and its co-founders of violating the lab’s founding agreement when it began taking major investments from Microsoft and making commercial products.
It’s asking for $150 billion and an order that pushes OpenAI to become a profitable company last year. Mr. Musk is also calling for Mr. Altman to be removed from the company’s board.
Judge Gonzalez Rogers’ Thursday morning instruction went straight to one of the main arguments in the AI race in the tech industry. For years, researchers and tech moguls have argued about the risks of artificial intelligence, not just to white-collar jobs like computer programming, but to humanity itself.
Outside of Silicon Valley, the debate may sound absurd. But in many tech companies, this has been the motivation for people who believe they are the only ones who can create AI safely. This has sparked an extraordinary boom, with projections of $900 billion this year for AI data centers, according to research by investment bank Evercore.
In an Oakland courtroom, Judge Gonzalez Rogers kept the debate to more grounded questions. She consistently told lawyers to focus on the facts about OpenAI and its founding, and asked them to stay away from tangential issues like drug use.
However, it was open season for OpenAI’s chief adviser, William Savitt, to question Mr Musk’s credibility.
Mr. Savitt played a video of Mr. Musk’s testimony from last fall that contradicted what he said in court on Wednesday. Last year Mr Musk said he had not read a key OpenAI document. But he acknowledged on Thursday that he had read at least part of the document, Mr. Savitt said.
It may have been a small point, but it was part of a strategy to discredit Mr. Musk. Mr. Savitt asked Mr. Musk if he ever directed the algorithm that drives X, his social media platform, to promote his own account. Mr. Musk said no.
There have been incidents where the company has made changes favor his account. In February 2023, shortly after buying the company, Mr. Musk became upset that a Super Bowl post by President Joseph R. Biden Jr. worked better than his. Mr. Musk demanded that his staff figure out what was wrong, and at one point made an adjustment that caused the site’s timeline to appear almost entirely of Mr. Musk’s posts. Later it switched back.
Russell Cohen, Microsoft’s general counsel, who was also named in Mr. Musk’s suit, raised questions along similar lines. He sought to show that Mr. Musk was well aware of the complex relationship between OpenAI and Microsoft as early as 2020, but decided to sue the tech giant only four years later.
Like OpenAI’s legal team, Mr. Cohen sought to demonstrate that Mr. Musk filed the lawsuit only after the stakes in the AI race had escalated, long after OpenAI’s ChatGPT had become a global phenomenon.
Mr. Molo, Mr. Musk’s lawyer, who opened Mr. Musk’s questioning on Tuesday, returned to ask the billionaire a few final questions. He focused on why Mr. Musk didn’t file a lawsuit years ago.
It’s an important point that Mr. Musk’s party wanted to make. He has long insisted that OpenAI should not be run by a for-profit company.
Mr. Musk, repeating what he said in earlier testimony, said his electric car maker Tesla has no plans to push AGI, or artificial general intelligence, essentially a machine that can do everything a human brain can do. That seems to contradict a recent social media post in which he said Tesla would be one of the companies building the technology.
When Mr. Musk unequivocally denied that Tesla was doing AGI, OpenAI’s Mr. Brockman scribbled a note on a small piece of yellow paper and handed it to his legal team.
As his testimony concluded, Mr. Musk still managed to slip in some references to the movie “Terminator,” the classic movie about an artificial intelligence that nearly destroys humanity.
As Mr Musk left the courtroom, he passed the head of his family office, Jared Birchall, who was next on the witness stand. They didn’t recognize each other.
Robert Kry, one of Mr. Musk’s lawyers, took Mr. Birchall through about 60 donations Mr. Musk made to OpenAI between 2016 and 2020, including rent payments for OpenAI’s San Francisco offices after Mr. Musk left its board. Mr. Kry tried to show that Mr. Musk has put a lot more money into OpenAI than anyone else.
Mr. Birchall also discussed a $97.4 billion offer by Mr. Musk and others to buy OpenAI’s assets last year. He said he was concerned that Mr. Altman was inappropriately taking away value from the nonprofit OpenAI when he and others created a new for-profit company in anticipation of a public offering.
The bid, Mr Birchall said, was designed to “prevent the foundation from being devalued” by establishing its market value.
Under cross-examination, Bradley Wilson, a lawyer for OpenAI, asked Mr. Birchall whether he had coordinated the $97.4 billion bid with one of the lawyers who filed Mr. Musk’s lawsuit against OpenAI. After Mr. Birchall acknowledged that he did, OpenAI’s legal team asked Judge Gonzalez Rogers to strike his testimony about the offer from the court record. She said she would make a decision on the matter on Friday.
(The New York Times has sued OpenAI and Microsoft, alleging copyright infringement of news content related to AI systems. Both companies have denied the suit’s claims.)
The trial is expected to resume on Monday.
Ryan Mac contributed reporting from Los Angeles.
