House Democratic Leader Hakeem Jeffries said Sunday that Congress must move "urgently" to address the risks of artificial intelligence, calling for decisive action to slow the pace of development — putting his party's leader in the House squarely behind intervention as the AI safety debate becomes a defining political fight heading into the midterms.
"We certainly need to begin to act urgently to address the challenges that exist," Jeffries said on ABC's "This Week with George Stephanopoulos." "We should take decisive action now so that we can slow down, as the CEOs have recently acknowledged, slow down the pace of development in order to protect the American people and ensure that AI is proceeding safely." For more context on this story, see our ongoing latest AI developments.
Prompted by Warnings From Inside the Labs
Jeffries' comments followed a stark public warning from Jacob Coxon, a former Anthropic researcher. In a thread on X, Coxon wrote, "The people building AI earnestly believe that it could kill us all by the end of the decade."
Appearing on "This Week" on Sunday, Coxon said that within tech companies, conversations reflecting those dire sentiments have been increasing in "recent months," and he urged the public not to leave the decisions to the people building the systems. "It's pretty important to keep your eyes open about the direction things are heading and not necessarily leave it all up to the people that are building it because, you know, they want engagement," Coxon said. "The people building the technology want to have conversations with people."
"One of the main sticking points over the last few years has been the inability to wake people up about the danger that they're creating," he added.
Democrats Prepare to Put AI on the Agenda
According to Jeffries, his caucus is ready to translate those warnings into legislative work. The Democratic caucus plans to meet on Tuesday, and the House leader said "a high priority will be action on artificial intelligence."
That timetable means concrete proposals could begin taking shape within days, giving Democrats a policy answer to a story that has dominated the news: AI executives themselves — including Anthropic's Dario Amodei and OpenAI's Sam Altman — publicly acknowledged in recent days that the technology is advancing faster than society can absorb, with some pleading for Washington to step in.
A Washington Split Over Who Acts First
The reaction to the industry's slowdown call has split along predictable but hardening lines. President Trump rejected the calls outright, dismissing the critics as "negative forces" and warning that slowing down would cede ground to China. House Speaker Mike Johnson, in a separate Sunday interview, argued Congress should resist any "emergency moratorium" and that tech companies must first take "personal responsibility" for the safety of their models.
Jeffries, alongside former President Barack Obama — who urged Democrats this weekend to move AI oversight to the center of their agenda — now represents the other pole: that Congress, not voluntary restraint by a handful of companies, should set the rules.
The Democratic leader's framing was notably careful. He tied the case for action to the executives' own admissions, effectively arguing that if the builders themselves say the pace is too fast, Washington has both the justification and the obligation to act. It is an argument designed to be difficult for opponents to dismiss: the premise comes not from activists, but from the CEOs of the labs doing the building.
From Warning to Legislation
What "decisive action" would look like in practice remains undefined. The safety debate has circulated a familiar menu of measures — mandatory evaluations before frontier releases, transparency about model capabilities, incident reporting, and federal commitments on pacing development. None of those specifics came from Jeffries on Sunday; what he offered was direction, not a bill. The task for his caucus when it meets Tuesday will be to turn that direction into legislative text.
What to Watch
Tuesday's caucus meeting is the first concrete marker. If AI action is truly a "high priority," expect policy frameworks — safety testing, transparency requirements, pacing commitments — to surface in the coming weeks as Democrats seek to convert public anxiety and industry testimony into legislation.
The counterpressure is equally organized: a White House and congressional GOP leadership opposed to rushed rules, and a geopolitical argument that slowing American labs benefits Chinese rivals. How that balance is struck may determine whether the United States gets AI legislation before the midterms — or whether the fastest-moving technology in history remains, for now, governed largely by the judgment of the people building it.
With control of Congress on the ballot in November, both parties now have incentives to stake out clear positions. AI, long a niche portfolio handled by subcommittees and think-tank panels, has abruptly become a question leaders are asked about on the Sunday shows — and, increasingly, one that voters may reward or punish them for answering.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
