WASHINGTON, Oct. 7, 2026 | OpenAI CEO Sam Altman said in an interview with Politico's Decoded newsletter, published Oct. 4, that the world "should accept some bad things happening for the benefits of this technology." He rejected a hypothetical trade in which AI would be made free of hacks and misuse at the cost of keeping powerful systems inside one lab. He also said some risks, including a "serious loss of control to AI," are not acceptable.
What Altman Said About Accepting Risk
Altman argued against making zero harm a condition of releasing AI broadly. SiliconANGLE quotes him: "I wouldn't take a trade of saying, 'we'll make sure there's no major hacks, there's no misuse...'" He said people will do "orders of magnitude more" good than bad with the technology, per Futurism and Breitbart's accounts of the interview.
He called the scenario of keeping powerful AI in a single lab while controlling its distribution a "completely unacceptable trade-off." Politico's interview was not available to me directly, so these quotes come from outlets that reported it.
Where He Draws the Line
Altman backs what outlets describe as lighter-touch regulation. He said oversight should focus on "really catastrophic risks," and named "serious loss of control to AI" as an example, while letting people use AI for beneficial purposes. Futurism reports that he said that if something goes wrong with OpenAI's models during training, "there's going to be some version of that we need to be responsible for."
Per SiliconANGLE, Altman said there is "a lot of daylight" between his view and that of Anthropic CEO Dario Amodei. ExplainX notes that his description of Anthropic's approach is his own framing of a rival's position, not Anthropic's stated view. Breitbart, summarizing the interview, says OpenAI has recently backed state safety requirements and bipartisan proposals for external safety evaluators.
Why the Timing Matters
The interview came as several governments weigh how to oversee frontier AI. New York City's council this week questioned AI companies under oath, as we covered in our NYC Council story, and the White House announced a voluntary accord and a new AI task force, which we covered in our Super Intelligence Force story. Altman's remarks are one executive's position, and Anthropic and other labs hold different views.
What This Means for Businesses Using AI
If broad access to AI models stays the default, as Altman argues, then more of the safety work falls on the companies building with them: abuse monitoring, access controls and clear responsibility for what an AI agent does. The practical question for a buyer is not which lab has the safest philosophy but what safeguards the vendor documents, and what the company has to add on its own.
Frequently Asked Questions
What did Sam Altman say about accepting AI risks?
Altman told Politico's Decoded that the world "should accept some bad things happening for the benefits of this technology," and said he would not trade broad access for a guarantee of no hacks or misuse. He also said catastrophic risks such as loss of control are unacceptable.
What is Altman's position on AI regulation?
Altman supports lighter-touch regulation focused on "really catastrophic risks," according to coverage of the interview. He rejected keeping powerful AI inside a single lab, calling it a "completely unacceptable trade-off."
How does Altman's view differ from Anthropic's?
Altman said there is "a lot of daylight" between his view and Dario Amodei's, per SiliconANGLE. His description of Anthropic's approach is his own framing, not Anthropic's stated position.