Anthropic CEO and 1,386 AI Experts Call for Slowdown in Frontier Development

Anthropic CEO Dario Amodei and a coalition of over 1,300 AI industry insiders are calling for a measured slowdown in the development of frontier AI capabilities. The push for "pacing" aims to ensure that safety protocols and operational excellence keep pace with the rapid evolution of artificial intelligence.

In a September 2026 essay titled "We Must Pace the Frontier," Amodei argued that slowing the rate of improvement would allow the industry to address critical risks currently overlooked due to the sheer speed of development. "We must slow the pace at which we improve the capabilities of AI models," Amodei wrote, noting that a more measured approach would enable "much greater operational excellence."

This sentiment is echoed in a July 2026 open letter signed by 1,386 employees from leading AI firms, including Google, OpenAI, and Anthropic. The signatories argue that the industry needs to "buy time" to develop essential security measures and governance frameworks. The experts warn that without such a pause, capability acceleration could soon exceed the human ability to understand or control these systems.

The AI Arms Race

The call for restraint comes amid a global "arms race" where companies and nations compete for economic dominance and national security advantages. This environment has fostered deep concerns regarding recursive self-improvement and the potential for "emergent malevolence" within AI systems. The current trajectory is viewed by some insiders as a dangerous gamble with global stability.

Elizabeth Edwards-Appell, an employee at Anthropic and a signatory of the letter, described the pursuit of advanced AI without guaranteed human control or international governance as "objectively, an insane and suicidal thing to do."

Industry Implications

That these warnings are coming from the engineers actually building the technology suggests a growing internal consensus that existing safety protocols are insufficient for the scale of capabilities being developed. Experts have compared the current risk profile to the early days of nuclear physics, where the fear of runaway reactions mirrored the current fear of losing control over autonomous systems.

Micah Carroll of OpenAI highlighted the geopolitical danger of this competition, stating, "I fear that in an international race to the bottom of AI development, it is likely that no nation will win, and we will all lose together."

The Path Forward

The movement now focuses on whether the industry can move from individual warnings to a coordinated international framework. The primary objective remains the establishment of guaranteed security measures that can prevent a catastrophic loss of control. While the technical capacity to accelerate continues to grow, the ability of the global community to implement binding safety standards remains the critical variable to watch.