Not long ago, most of Washington was scrambling to be as friendly as possible to the rise of artificial intelligence. Lawmakers from both parties saw a chance to bolster the nation’s economy—and their campaign coffers. But rarely has public opinion flipped so quickly and so thoroughly as it has on AI in recent months. While AI was making some tech leaders unfathomably rich, many Americans in states both red and blue grew alarmed by the proliferation of loud, electricity-guzzling data centers, convinced that AI would first make life more expensive and then, for good measure, eliminate their job. In the past few days, some of the top voices in the AI industry have gone further, warning that unless development of the technology is slowed, and slowed soon, all of humanity could be made extinct.

These fears have rattled markets and galvanized a bipartisan set of leaders into calling for the government to step in. AI has now become a central issue in the midterm elections. But President Trump, who first came to power by closely reading the mood of the country and shifting his views to meet those of the electorate, seems to have his fingers in his ears, ignoring it all and continuing to embrace AI, confident that everything will be fine because he is in charge. Unless, of course, the human race ends first.

Trump would seem to have a perfect opportunity to try to contain this fast-moving political conflagration: Next week, he’s set to host China’s president, Xi Jinping, at the White House. The United States and China are the world’s leaders on AI and are in a race to develop the technology. The summit could be a venue, some in the industry hope, to order a pause in development or establish international safety guidelines. But Trump, at least for now, wants no part of a slowdown. AI growth has propped up Wall Street and sectors of the American economy for the duration of his second term, and the president has made clear that he doesn’t want to do anything that could burst that bubble.

“The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!” Trump wrote on Truth Social today, as he made half a dozen posts, each seemingly more frantic than the last. He later compared the concerns about the technology to his own impeachments, declaring that “AI taking over the World, destroying Humanity, and all other things bad, is a HOAX.”

As tech insiders offered warnings about the AI apocalypse, Trump was busy rallying the MAGA faithful in Texas and taking in rounds of golf in Ireland. The dire prophecies began last Tuesday, with a series of viral social-media posts by an AI researcher who said that the technology is advancing so quickly that it could bring about human extinction by the end of the decade. The threat of annihilation seemed to grab Washington’s attention as few other concerns have. By the next afternoon, after several other researchers had echoed the alarm, members of Congress were beginning to call for legislation, new agencies, hearings, and regulations to address the threat. Senator Bernie Sanders of Vermont told us on Wednesday that he had sent invitations to a bipartisan set of lawmakers for a closed-door briefing with AI experts. He said that the researchers had provided a wake-up call about a technology that poses an “extraordinary danger for the human race.” That night, Trump spoke for nearly two hours at the Republican midterm convention in Dallas, but he did not mention AI once.

Read: AI is running laps around our political system

In the days that followed, more troubling evidence emerged. Anthropic released a report detailing the ways that individuals and nation-states have attempted to misuse its technology, including for sophisticated cyberattacks and research that could potentially support the development of biological weapons.

Trump decamped to Ireland, where he spent the weekend watching a golf tournament at one of his properties. He was dismissive of the Silicon Valley warnings, emphasizing the need to defeat China in the AI race and downplaying the technology’s risks. On Saturday, Anthropic CEO Dario Amodei wrote that a downshift in the pace of AI development “could greatly reduce the risk that something goes seriously wrong.” Other industry leaders joined in, including OpenAI’s Sam Altman and Elon Musk, the sometimes–Trump ally.

One of Trump’s first moves upon returning to office last year was to revoke the executive order signed by his predecessor, Joe Biden, that imposed a series of regulations on the AI industry. Trump’s focus was business growth, and he reveled in the power he had over the nation’s tech titans. Many crowded the room at his inauguration, and he has hosted AI-industry leaders such as Altman, former Oracle CEO Safra Catz, and Nvidia President Jensen Huang.

Still, some of Trump’s advisers have had concerns since last year about the political impacts of the AI backlash, when figures as disparate as Sanders and Florida Governor Ron DeSantis began calling for more regulation. To head off the potential populist backlash, Trump announced the Ratepayer Protection Pledge in March, asking major developers of AI to “build, bring, or buy” resources needed for any new electricity-production and utility-infrastructure upgrades. Now the White House is working with House Speaker Mike Johnson to fast-track legislation that would put into law some of the pledge’s voluntary measures. A vote is scheduled for tomorrow.

From the March issue: America isn’t ready for what AI will do to jobs

People close to the administration perceive some figures, including White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent, to be supportive of additional regulation of the industry, two people familiar with internal White House deliberations told us, speaking on the condition of anonymity. Others, such as Commerce Secretary Howard Lutnick, have advocated for as little government interference as possible, while Defense Secretary Pete Hegseth has demanded that Anthropic abandon ethical safeguards—such as rules against mass surveillance and autonomous lethal targeting—in its work for the Pentagon. (When Anthropic refused, Hegseth cut ties with the company.)

Trump has dug in on letting AI rip. He has watched as the industry has propelled the stock market, the president’s favorite indicator of a strong economy, keeping it high even amid challenges posed by the Iran conflict, stubborn inflation, and tariffs. Trump has expressed worry to those close to him that a halt to AI expansion could send the economy into a recession, one ally told us.

Some in the president’s orbit have raised suspicions about the AI executives’ warnings: Is this an attempt at market capture? (They worry that new regulations would freeze the industry and prevent would-be competitors from catching up to the industry giants.) Vice President Vance seemed to give expression to that idea by saying today that tech companies “begging the government to regulate them” feels “like a bit of a Trojan horse.” David Sacks, a co-chair of the President’s Council of Advisors on Science and Technology, wondered on social media over the weekend why AI companies cannot just regulate themselves; he told tech leaders to “stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel.” (Top AI firms could work together to implement some safety standards, but antitrust law would likely prevent them from implementing any kind of reciprocal agreement to throttle the pace of development, unless Congress granted a waiver for that purpose, MacKenzie Arnold, the managing director of the Institute for Law and AI, a nonpartisan think tank, told The Atlantic.) Today, Trump called Huang while the CEO was onstage at an AI event to assure the Nvidia head that he will keep supporting the technology.

Read: J. D. Vance’s AI doctrine

The pleas for action have created unlikely political bedfellows. Senator Ted Cruz of Texas, a Republican, declared that “we need substantially more regulation,” while Senator Brian Schatz of Hawaii, a Democrat, said that Congress needs to “move at the pace of the threat.” Tomorrow, Sanders will join the Trump adviser Steve Bannon at a “Pro-Human Assembly,” an all-day gathering in Washington, D.C., focused on ensuring that humans remain in charge of AI. Despite support for regulation from both the right and the left, there is little meaningful oversight of AI that can be implemented without the president’s approval. Sanders told us that he hopes the bipartisan outrage will ultimately get Trump’s attention. “Trump’s ignorance is extremely dangerous,” he said. “He has not a clue about the dangers inherent in this technology, nor do people in his administration.”

Trump is looking forward to next week’s summit with Xi, which follows a May meeting in Beijing. The U.S. president has stressed his strong relationship with the Chinese leader and has downplayed China’s efforts to thwart the United States. It seems unlikely that Trump will let AI fears fully derail the summit’s agenda. Trump has also minimized China’s chances of catching up to the U.S.—he green-lighted a sale of Nvidia chips to China late last year—even as he has played up the stakes. “Whoever wins AI wins,” he said this past weekend.

But Xi may come into the meeting with a new wariness of his own. China’s top intelligence official warned yesterday that AI could pose a direct threat to the Chinese Communist Party’s hold on power by enabling sophisticated hacking and disinformation campaigns. Many in the industry hope that the summit will be an inflection point. Nate Persily, a co-director of the Stanford Law AI Initiative, told us that “if ever there was an opportunity to get some kind of consensus on AI safety, this provides that opportunity.” And Altman, in an interview over the weekend, made a clever appeal to Trump’s vanity, suggesting that if he and Xi strike some sort of deal to create AI guardrails, the president could win his long-coveted Nobel Peace Prize.

Michael Scherer, Ashley Parker, Will Oremus, and Jack Rodriquez-Vars contributed reporting.