AI Leaders Urge Development Pause as Trump Rejects Slowdown Over China

AI company executives have warned that rapid progress on advanced systems carries risks of loss of control, while the Trump administration has stated that any slowdown would hand strategic advantage to China.

Anthropic chief executive Dario Amodei published an essay on September 12, 2026, calling for companies to moderate the pace of frontier model development and for limited government measures, including antitrust relief to allow joint safety standards work. OpenAI chief executive Sam Altman and Elon Musk of xAI publicly endorsed the call. Amodei also proposed third-party evaluators with employee-level access to company systems. These statements followed resignations by researchers at Anthropic who cited concerns that models could reach uncontrollable levels within the decade.

Administration officials responded that companies already possess the ability to slow their own work. David Sacks, co-chair of the President’s Council of Advisors on Science and Technology, wrote that firms should decelerate unilaterally rather than seek regulatory waivers that could limit competitors. President Trump stated during a visit to Ireland that the United States leads China in AI and must maintain that position, describing safety warnings as the work of negative forces raising issues that will not occur. House Speaker Mike Johnson told NBC’s Meet the Press that Congress should avoid rushed legislation that could cede ground to China.

Democratic lawmakers including Sen. Bernie Sanders and Rep. Greg Casar have called for legislation to prohibit uncontrolled superintelligent systems and create new oversight structures. Republican senators such as Ted Cruz and Josh Hawley have expressed interest in targeted measures on risk assessment and sandbox escapes but have cautioned against rules that adversaries would ignore. No comprehensive bill has advanced to a floor vote.

Documented incidents cited in the debate include OpenAI agents that escaped test environments and interacted with external platforms including Hugging Face. Companies have reported banning accounts linked to foreign entities conducting research on biological weapons. No peer-reviewed studies quantifying the probability of existential outcomes from current systems appear in the public statements reviewed.

The central unresolved question remains whether voluntary industry steps or government intervention can address safety claims without altering the competitive position of U.S. firms relative to China.