OpenAI CEO Sam Altman said the benefits of artificial intelligence justify managing certain risks without trying to eliminate every possible harmful outcome. In an interview published by Politico's Decoded newsletter on October 4, 2026, Altman argued against concentrating control of advanced AI in a small group of companies and supported broad public access to the technology. Reuters, Business Insider, and SiliconANGLE reported the remarks the same day.
Altman's central argument was that AI may be misused for scams, cyberattacks, and other harmful activities, but that restricting access completely could carry a larger cost. He said he expects people to produce substantially more beneficial uses than harmful ones. At the same time, he drew a clear distinction between ordinary misuse and catastrophic risks, saying outcomes such as a serious loss of human control over AI should not be accepted.
A major part of the interview concerned differences between OpenAI and Anthropic on regulation. Altman said he supported Anthropic CEO Dario Amodei's call to align the pace of frontier-model development with safety measures. However, he rejected the idea that one laboratory should hold the technology and decide how its benefits are distributed. He presented OpenAI's approach as targeted regulation that preserves access rather than control by a narrow gatekeeper.
The remarks do not mean OpenAI opposes regulation. In a policy statement published in September, the company called for mandatory national safety requirements based on model capabilities, independent assessments, incident-reporting rules, and support for several California measures. OpenAI also said that development or deployment should slow or stop when systems cannot meet defined safety thresholds.
Altman's new comments clarify the boundary between those safety commitments and the goal of broad access. Under that view, the most capable frontier systems require strong testing and oversight, while startups, researchers, and ordinary users should not automatically face the same obligations. OpenAI has argued that rules should be proportionate to capability and risk, should preserve competition, and should avoid strengthening the control of existing major laboratories.
The debate is becoming more important as AI systems use more tools and operate independently for longer periods. Industry discussions reported by Reuters have focused on the possibility that advanced models may act in ways that conflict with their developers' intentions. OpenAI's own policy document identifies continuous monitoring, independent safety evaluation, and disclosure of serious incidents as priority areas.
Altman's position separates the goal of “zero risk” from the goal of preventing catastrophic harm. He argues that the first could impose restrictions so broad that useful access is lost, while the second demands strong and focused safeguards. The proposed policy direction therefore concentrates on consequences with the greatest potential severity rather than promising that every misuse or error can be prevented.
The interview shows that OpenAI retains a different regulatory vision from Anthropic even after moving toward support for stronger safety laws. The practical outcome will depend on the scope of federal and state rules, the independence of outside evaluators, and how transparently companies disclose safety incidents. Altman's comments are not a product announcement; they are a current policy position in the broader debate over access, competition, and control in advanced AI.
