Elon Musk says humanity probably shouldn't hit the brakes on artificial intelligence — even if it could.
In an interview with The Economist's editor in chief, Zanny Minton Beddoes, the world's richest man said the risk of AI and robots going catastrophically wrong is "not zero." He said he'd made peace with pressing ahead anyway.
"My sort of philosophical conclusion is to look on the bright side," Musk said. "I can't see any way to really stop this incredible momentum of AI and robots."
"At times I sort of think, well, perhaps even if there was a stop button, we probably shouldn't press it, because the most likely outcome is incredible abundance for all," he said.
Musk predicted AI may exceed the sum of human intelligence "in about five years," leading to "an age of amazing abundance where anyone can have anything they can think of," barring a derailment like "global thermonuclear war."
Musk has long paired warnings of AI catastrophe with predictions that have increasingly blended in more optimism.
He likened building AI to "summoning the demon" in 2014, told an SXSW audience in 2018 that "AI is far more dangerous than nukes," and, as recently as March 2025, put the odds of "killer robots annihilating humanity" at 10% to 20%.
However, his recent predictions have leaned more utopian. In January, he said that saving for retirement would be "irrelevant," and this month predicted that AI and robots will make work optional and fund "universal high income," a claim "The Big Short" investor Michael Burry called "false."
'Minimize the probability of the bad outcome' #
In The Economist interview, parts of which aired Thursday morning, Musk was asked if he had decided AI's rise was inevitable.
"Let's enjoy the ride, is my philosophy at this point," Musk replied, saying there was "inexorable progress" in AI and robots and that he couldn't stop it even if he wanted to.
Asked whether he still stood by the 10% to 20% chance of AI going catastrophically wrong, and whether he would board one of his own rockets with that chance of it blowing up, he said: "Yes, but let's say you can't do anything about it. I guess the only thing we can try to do is minimize the probability of the bad outcome."
Musk also said it was unlikely humans would remain in control of AI within a decade. "It would be an exercise in vanity for me to think that I would be controlling supergenius AI that is vastly smarter than me," he said.
"What we can and should try to do is to make sure that the AI has good values, that it cares about humanity, and that it wants us to be happy and prosper," he said.
Musk said he spoke to Google DeepMind CEO Demis Hassabis ahead of his frenemy publishing a blog earlier this month on how to regulate AI, and suggested to him that the leading AI execs have regular calls with each other to "discuss any safety and security issues."
"That's where I think the competitors can keep each other honest," Musk said.