A Nation on Edge: Two‑Thirds of Americans Fear AI Could Bring Humanity’s End

When you ask strangers on the street what worries them most, the answer used to be the economy, climate or health care. Today, for many Americans, it’s something that sounds like science‑fiction: artificial intelligence that could wipe us out.

A new Politico poll that surveyed a little over 2,000 adults this September found that roughly two‑thirds of respondents think there’s at least a “moderate” chance that advanced AI could end humanity. That’s 66 % of the country looking over its shoulder, and the fear isn’t confined to any single political camp.

Even among those who voted for Kamala Harris in 2024, 70 % said they view AI as an existential risk, while 60 % of Trump voters shared the same concern. The numbers are soberingly close, suggesting that the dread of runaway machines cuts across the usual partisan divide.

What’s perhaps more unsettling is the share of people who are almost certain that AI will be our undoing. Seventeen percent of the surveyed adults answered “almost certain,” and that sentiment split evenly between Harris and Trump supporters. By contrast, only 22 % said the technology poses a very small or no risk at all.

These figures arrive at a moment when the AI community itself is sounding the alarm. Last week an Anthropic researcher quit the firm, accusing the company of “gambling with our lives,” and claiming the engineers building these systems genuinely believe they could kill us all. Around the same time, Anthropic co‑founder Dario Amodei published a long‑form essay urging the United States’ biggest AI players to collectively slow down development—a plea that earned public backing from Elon Musk and OpenAI’s Sam Altman.

Still, not everyone buys the industry’s safety narratives. The poll found that only about a quarter of respondents—26 %—think the warnings about existential risk are just a marketing gimmick. The majority, however, seem to take the warnings seriously: 48 % of Americans say AI has advanced enough that we should hit the pause button entirely.

Even here, party lines shift a bit. Forty‑four percent of Trump voters favor a full pause, compared with 58 % of Harris supporters. Conversely, a larger slice of Trump backers—40 %—think we should keep pushing forward because of the potential benefits, versus 28 % of Harris voters.

So the picture is clear: most Americans see AI as a genuine threat, but the nation is split on how to respond. Whether policymakers heed the public’s unease—or whether tech companies decide that the promise outweighs the peril—remains to be seen. One thing is certain, though: the conversation about artificial intelligence has moved from the lab into living rooms across the country, and the stakes feel higher than ever.