Sam Altman Says Some AI Harm Is the Price of Progress
OpenAI chief executive Sam Altman said the world should accept some harm from AI in return for its benefits.
He named hacks, scams and misuse as problems that will happen, according to The Guardian. He favors lighter-touch rules, where society builds resilience after tools are in use. Reuters quoted him saying: "We believe that the world should accept some bad things happening for the benefits of this technology and people having the agency..." in its account of the interview. The Guardian links the remarks to Politico's Decoded podcast released on Monday, published as the Decoded interview.
The comments followed a White House pact last week signed by Donald Trump and AI bosses including Altman and Dario Amodei of Anthropic. Under that deal, Trump said companies should self-police to limit risks from AI-enabled cyber-attacks and bioweapons.
Florida pushed back. Governor Ron DeSantis said he had "no dice" with a handful of tech oligarchs deciding safety for others. The State of Florida asked a judge to bar OpenAI from developing new AI models without third-party approved guardrails, or outside-reviewed safety checks.
The debate coincided with internal dissent and public worry. An OpenAI safety leader quit, warning the company's culture was broken and AI firms were not being nearly careful enough, according to The Guardian's Oct. 3 report. A separate Reuters report framed the warning as a claim that the time for trial and error was over. Three-quarters of Americans worry AI firms have not gone far enough to prevent serious harm, Reuters reported.
Altman has opposed pre-release controls. He planned to urge U.S. lawmakers against requiring developers to get government approval before releasing models, Reuters reported in June. He said OpenAI will not hold an IPO in 2026 and feels no pressure to go public. He called a 10% risk of AI-caused extinction "unacceptable," according to Reuters reporting from September.
The broader context here is a clash of regulatory ideas. Washington is testing voluntary, company-led policing for catastrophic uses. Florida is testing court-ordered, third-party checks as a condition for development. Congress is being asked whether pre-release government approval should exist at all.
For what to watch in governance, the test will be practical. Altman assumes harms like fraud and intrusions can be absorbed while defenses improve. Critics, including state challengers and departing safety staff, assume the window for trial and error has narrowed. Polls give elected officials reason to press for harder guarantees. The next markers are whether a court considers a development freeze tied to outside review, and whether self-policing pledges gain enforceable standards or stay voluntary.
