More Than 20 Lawmakers Demand AI Regulation After Researcher's Extinction Warning, But Congress Is on Recess Until the Midterms

Since former Anthropic researcher Jacob Coxon announced his resignation Tuesday, warning that AI could cause human extinction by the end of the decade, more than 20 members of Congress have publicly called for new or stronger AI regulation, according to CNBC.

Coxon, who spent three years doing pretraining research at both OpenAI and Anthropic, posted on X that "the people building AI earnestly believe that it could kill us all by the end of the decade" and accused both companies of "racing straight to self-improving superintelligence and gambling with our lives." The post has drawn more than 150 million views, per CNBC.

Evan Hubinger, Anthropic's alignment science lead, backed Coxon's assessment publicly, putting his estimate of the risk above 10% over the next decade, according to Quartz. Another Anthropic employee wrote, "We really do earnestly believe AI could kill all humans!" as reported by The Guardian.

Lawmakers React Across Party Lines

Rep. Lori Trahan, D-Mass., told CNBC's "Squawk Box" on Friday that bipartisan support has hit "a tipping point." "My phone has rung off the hook this week with rank-and-file Democrats and Republicans who want to see action," she said.

Rep. Nathaniel Moran, R-Texas, posted on X that "innovation and safety are not mutually exclusive" and that Congress "cannot ignore the realities of AI and the potential risks that come with it." Rep. Ted Lieu, D-Calif., called Coxon's warning "exhibit number 739" for why the AI Kill Switch bill needs to pass, per The Guardian.

Sen. Ted Cruz, R-Texas, told ABC's "The View" that AI poses a "catastrophic risk," recalling that Elon Musk once told him on his podcast the odds of AI destroying humanity were "10-20%." Sen. Bernie Sanders, I-Vt., said on X that "Mr. Coxon is right" and that he plans to introduce legislation to "ban superintelligence and pause AI development."

Sen. Ruben Gallego, D-Ariz., wrote to Senate Majority Leader John Thune, R-S.D., and Minority Leader Chuck Schumer, D-N.Y., asking them to form a bipartisan Senate Select Committee on Artificial Intelligence with subpoena power, arguing that "jurisdiction over AI is scattered across multiple Senate committees, each with a focused lens but none with the full picture," according to Quartz.

Three Bills, No Floor Votes

Congress already has AI legislation sitting on the table. Reps. Trahan and Jay Obernolte, R-Calif., introduced the FRONTIER Act in July to set a framework for deploying advanced AI models. The same day, Moran and Lieu introduced the AI Kill Switch Act, which would require companies to maintain the ability to shut down or throttle their systems. Earlier this month, Sanders and Rep. Greg Casar, D-Texas, unveiled the Ban Artificial Superintelligence Act, which would pause advanced AI development until federal safety rules exist.

None of the three bills has advanced past introduction, according to CNBC.

Public Unease Is Rising, Political Will Is Not

A Pew Research Center survey of 3,488 adults conducted June 22-28, with a margin of error of plus or minus 1.8 points, found 52% of Americans are more concerned than excited about AI's growing role in daily life, up from 37% in 2021.

Despite that, Axios reported Congress is at the "concepts of a plan" stage, with House Republican leadership, which controls both chambers, showing no indication it views this as urgent. Axios also reported that many Democrats are treating AI regulation as a 2027 project, contingent on winning back a chamber majority in the midterms. Both the House and Senate are largely in recess ahead of November, per CNBC and Quartz, leaving little runway for any bill to reach a vote this year.

President Trump dismissed extinction concerns Thursday, telling reporters, "if we don't win AI, we're going to be put in a very bad position," according to Quartz. His administration has continued to prioritize AI expansion, including a recently reported deal to equip federal employees with discounted access to OpenAI's models, per Fox News Digital.

The Unresolved Question

The dispute isn't whether AI companies' own researchers are worried. Hubinger and Coxon, both insiders, put real numbers on the risk. What's contested is whether that internal alarm translates into an imminent, provable threat or reflects subjective forecasting under deep uncertainty about how future systems will behave. No government agency, court, or independent scientific body has established a mechanism by which current AI systems could cause mass casualties, and the estimates cited by Cruz, Sanders, and Hubinger are personal probability judgments, not measured outcomes.

What is measurable: three bills sit stalled in a Congress that's mostly out of session. No committee vote is scheduled on any of them, and the administration is simultaneously expanding federal AI adoption while dismissing extinction talk as a distraction from the competitive race with China. Whether any of that changes depends on what happens when lawmakers return from the midterm campaign trail and whether voter anxiety, now sitting at 52% in Pew's polling, follows them back to Washington.