OpenAI CEO warns of loss of control and power concentration as AI advances

OpenAI chief executive Sam Altman posted on X that rapid AI progress poses two major risks: loss of human control over increasingly powerful systems and excessive concentration of AI power in the hands of a single entity. He said alignment and safety work must stay ahead of capability gains and warned that a powerful AI used by one person or company to impose a worldview could lead to a dystopian outcome. Altman urged the industry to adopt consistent safety standards and supported a federal framework for frontier AI, while emphasizing that companies should act responsibly without waiting for government action. He noted that OpenAI now prepares explicit safety cases before major reinforcement-learning runs and called for shared industry protocols on misalignment, monitoring and safety, acknowledging that such measures may slow development but are justified. Altman added that a federal role will eventually be needed to coordinate international safeguards.

Anthropic CEO Dario Amodei also called for slowing the pace of frontier AI to allow safety measures to keep up, proposing stricter testing of new models and independent evaluator access similar to a food-inspector system. He expressed support for federal regulation and a possible “kill switch” for dangerous systems, while rejecting an outright ban on advanced AI.

AI policy expert Neil Chilson criticized the industry leaders’ appeals for slower development, arguing that government restrictions could harm U.S. competitiveness against China. He described the calls as an attempt to position certain firms as the only ones capable of controlling the technology and warned against anti-competitive practices.

China’s foreign ministry responded to Amodei’s remarks about a Chinese lead in AI, calling fear-mongering and confrontation counterproductive to global AI governance and urging cooperation instead of conflict.