Sam Altman Says the World Is Right to Be Afraid of AI, Then Argues His Industry Should Handle It
The OpenAI chief acknowledged real danger at a San Francisco conference, then made the case for self-regulation. Politicians and some of his own peers said no.
Key points
- Sam Altman said Tuesday that the world is "right to be afraid" of AI, his first public comments since a viral post by a former Anthropic researcher warned AI could kill all humans by the end of the decade.
- Altman argued AI companies can self-regulate, saying he is "very confident" the industry will keep safety ahead of capability, or slow down if it cannot.
- Senator Bernie Sanders and former White House strategist Steve Bannon both called for government oversight, not industry pledges.
- Anthropic CEO Dario Amodei, who had already called on governments to regulate AI, said his company is now in "a dialogue with the rest of the industry" about binding safety standards.
- OpenAI, Anthropic and Google DeepMind are building a voluntary frontier AI standards effort with no published timeline or enforcement mechanism.
Sam Altman walked into a room of hundreds of business executives at a Salesforce conference in San Francisco on Tuesday and said something most tech bosses avoid: his industry genuinely frightens him.
"It doesn't take as much imagination as it used to to imagine how this could go wrong," he said. "I think the world is right to be afraid of this."
Then came the part his critics seized on. Altman argued that fear is not a reason for government rules. His company and others like it, he said, are capable of policing themselves. "We will get it right, I'm very confident in our company's and industry's ability to do this safely."
What actually triggered this?
A post by a researcher who recently left Anthropic went viral last week, claiming AI could wipe out humanity by 2030 if development continues unchecked. Several senior AI figures said they agreed, without explaining how they reached that conclusion. Washington noticed.
Our earlier report on Altman and Musk backing Amodei's slowdown call showed how quickly rival CEOs can close ranks around an alarm, and how fast that unity dissolves when regulation enters the room. Tuesday's conference landed squarely in that gap.
Nvidia chief Jensen Huang, whose company is the world's most valuable by market capitalisation thanks to demand for AI chips, backed Altman bluntly: "We don't need new laws or regulations. Safety is an engineering problem."
Huang's position is worth tracking. We noted on 17 September that he stayed silent when President Trump dismissed AI safety fears onstage at the All-In Summit, in our report on that exchange. On Tuesday he found his voice, and used it to oppose oversight.
Meta's Mark Zuckerberg posted on X that every lab already has financial incentive to build safely, because labs that cause harm face serious legal liability.
Should patients and ordinary people take the self-regulation promise seriously?
Not everyone in the room agreed. Jack Clark, an Anthropic co-founder, told the BBC that leaving AI as a "totally unregulated industry" was "rolling dice with immense risks." Patrick Hillman, chief operating officer of Logical Intelligence, put it plainly: "The only institution that Americans might trust less than Washington these days is Silicon Valley."
Hillman's challenge is the sharpest thing said all week. "If you believe what you are building is dangerous, show us what you are prepared to stop doing."
Senator Bernie Sanders argued that decisions over AI have been made by "a handful of the richest people in the world" and that the public, not oligarchs, should set the rules. Bannon was equally blunt: "They're disingenuous and they lie."
Amodei, appearing briefly at the same conference, said Anthropic is working with other labs on shared safety commitments. OpenAI executive Chris Lehane confirmed last week that the effort is explicitly voluntary.
Voluntary is the word to watch. The pledge exists. The teeth don't.
