Anthropic CEO Calls for AI Slowdown as Congress Struggles to Act

The CEO of Anthropic has publicly called for the artificial intelligence industry to decelerate its development pace, even as bipartisan efforts to regulate the technology remain mired in congressional gridlock.

Dario Amodei wrote in an essay published September 12 that the industry must "slow the pace" of AI advancement, according to reporting by The Atlantic. His statement comes amid a cascade of warnings from current and former AI researchers, including Jacob Coxon, who resigned from Anthropic over safety concerns and wrote that industry insiders "earnestly believe that it could kill us all by the end of the decade."

Why it matters

The disconnect between AI's acceleration and political systems' capacity to govern it represents a fundamental challenge for democratic institutions. When the people building transformative technology publicly warn of catastrophic risks while elected officials struggle to pass basic guardrails, the question shifts from whether regulation is needed to whether it's already too late.

Bipartisan proposals meet partisan reality

Lawmakers from both parties have introduced AI-focused legislation in recent weeks. Senator Ted Cruz and Senator Amy Klobuchar are developing a bill to prevent AI-induced biological or nuclear catastrophes. Representatives Josh Gottheimer and Mike Lawler proposed standards for deploying AI agents securely.

Yet none of these measures appear close to becoming law, The Atlantic reported. The political environment—characterized by election-season posturing and deep partisan divisions—has made substantive cooperation difficult. President Trump has rejected calls to restrict AI development, saying he wants to preserve America's technological lead over China. His position creates a significant obstacle for Republicans who might otherwise support stronger oversight.

Industry calls for its own regulation

Anthropic reported this week that it intervened to stop scientists from using its models to potentially develop biological weapons. Amodei warned that within six to twelve months, AI models could be "capable of taking over the entire internet with a persistent botnet," potentially causing hundreds of billions of dollars in damage.

OpenAI's chief global affairs officer, Chris Lehane, wrote that the U.S. faces a narrow window to establish guardrails "before AI capabilities outpace the institutions responsible for governing them." The company now supports state regulations it had opposed months earlier.

In July, approximately 1,400 AI company employees signed an open letter requesting government regulation of their industry.

Data centers become campaign flashpoints

Opposition to data centers—the physical infrastructure enabling AI advancement—has emerged as a potent political issue. In Wisconsin and Ohio, gubernatorial and Senate candidates are running attack ads branding opponents as pro-data-center. Senator Bernie Moreno of Ohio proposed a 100 percent federal tax on state or local incentives for data centers, calling the industry's pitch "the worst sales pitch in the history of sales pitches."

A Pew Research Center poll from August found that just over half of Americans feel more concerned than excited about increased AI use in daily life, up from 37 percent in 2021.

Senator Bernie Sanders told The Atlantic he organized a closed-door congressional briefing for this week with AI safety experts, calling Coxon's viral resignation post "an extraordinary wake-up call" for lawmakers. Sanders said he hopes politicians will work across party lines "out of a sense of self-preservation."

Nate Persily, co-director of the Stanford Law AI Initiative, said the rapid pace of developments makes it "very difficult to keep up with what the harms are and how to address them," noting that "there are genuinely hard questions of statutory design here."

The details were first reported by Toluse Olorunnipa in The Atlantic.