Altman warns AI could go badly if control is lost or power concentrates, urges pacing and safeguards

OpenAI CEO Sam Altman warns that rapid AI progress could go “very badly” in two ways—loss of human control and excessive concentration of power—and argues for a measured pace, stronger safety standards, and transparent safety evaluations. He advocates interim safety “cases,” industry-wide standards, and federal coordination, while noting that progress should slow rather than stop and that American companies should lead with responsible development and independent verification.