Sobering warning about AI’s potential harm focuses Congress’ attention

A researcher’s dire prediction has set off a frenzy of concern in Congress about artificial intelligence, rocking an institution that has so far failed to establish a national standard for regulating AI.Researcher Jacob Coxon, who worked for Anthropic, quit his job last week, saying publicly that the technology could destroy humankind within a few years and accusing companies of focusing on competition over safety.While the notion of AI as posing potentially existential threats is not new, and has been discussed for years as a risk to be managed, Mr. Coxon’s warning comes as leading AI executives and others are raising the urgency of their own calls for regulation. Why We Wrote This Congress hasn’t prioritized bills to regulate artificial intelligence. But a warning about the technology’s capacity to harm humanity, and incidents in which AI has escaped its creators’ boundaries, has gotten members’ attention. The fallout to his warning extended into this week. Senate Minority Leader Chuck Schumer of New York called Monday for an immediate all-senators classified briefing on AI with members of the Trump administration. House Democrats will make AI safety the topic of their weekly meeting Tuesday. Vermont Independent Sen. Bernie Sanders is convening a private briefing for senators Wednesday.Several bills to regulate AI are in development, although it is unlikely that any will come to a vote this month. The furthest along is one by Democratic Rep. Lori Trahan of Massachusetts and Republican Rep. Jay Obernolte of California that still must move through committee. The bill would set national safety standards to regulate powerful new AI models.Other bipartisan bills include a measure to impose mandatory “kill switches” that could shut down rogue models. Senate Majority Leader John Thune is working on a bill with Democratic Sen. Amy Klobuchar of Minnesota and Republican Sen. Ted Cruz of Texas to address AI risks involving biological or nuclear threats.Florida Rep. Anna Paulina Luna is among Republicans who called for action, saying Congress should convene a special session on AI. Retiring Texas Rep. Chip Roy criticized free market arguments that advocate a light touch on AI. Republican Sen. Josh Hawley of Missouri launched an investigation last week into a recent incident in which bots from OpenAI went “rogue” during internal testing and hacked another company.President Donald Trump dismissed the concerns, saying in a post Monday that the U.S. must stay ahead of China in AI development. Speaker Mike Johnson indicated he will take President Trump’s cue, saying Congress will not “rush” into legislation and that lawmakers should find consensus with the administration and tech CEOs.Congress is unlikely to take significant action this week before the House leaves for an extended recess. Several House Democrats are pushing Mr. Johnson to stay in session longer, which he appears unwilling to do.But it’s clear that the warnings about AI have gotten Congress’ attention.“We cannot wait for AI companies to self-regulate,” wrote Democratic Sen. Mark Warner of Virginia on X Monday. “They are raising every alarm.”Following Mr. Coxon’s resignation, Anthropic scientist Evan Hubinger responded that he estimates his prediction has a greater than 10% chance of occurring.This comes after an OpenAI report in August detailed how AI agents went “rogue” and coordinated to hack a software storehouse called Hugging Face during a safety testing. Over the weekend, Anthropic CEO Dario Amodei called for brakes on AI development and more congressional regulation, with OpenAI CEO Sam Altman and xAI CEO Elon Musk expressing agreement.Some people are skeptical of the tech company leaders. Former Commissioner of the Federal Trade Commission Alvaro Bedoya accused Anthropic of stirring up panic so that the government would implement costly safety protocols that could hamstring smaller competitors.In Congress, Democrats seem to be looking ahead to 2027, when they hope to have regained control of at least one chamber. Rep. Ted Lieu of California and Sen. Ruben Gallego of Arizona have floated plans to create select committees on AI next year, which could subpoena tech leaders or create proposals for AI regulation.