True
Sixty-three percent of Americans say AI could bring about the destruction of humanity, while 48 percent support halting AI development altogether, according to survey figures verified by this newsroom through three partner news outlets after the original Crosstab data document could not be reviewed directly.
The two findings sketch a public that is deeply uneasy about where the technology is headed. A 63 percent share saying AI could end humanity places existential concern well above the halfway mark among U.S. respondents. The 48 percent who would back stopping development sit just short of a majority โ a measure of how far support for a pause extends beyond researchers and industry critics to ordinary Americans.
U.S. attitudes toward AI risk feed directly into the regulatory climate surrounding frontier LLM development and the semiconductor capacity behind it, from memory supply to foundry output supporting GPU data centers. Polling of this kind is closely watched by Korean AI and semiconductor players such as Samsung Electronics, SK hynix and Naver for what it signals about the direction of U.S. oversight debates.
Reviewers examined 10 primary sources. The original Crosstab source file in PDF form could not be confirmed, so verification relied instead on cross-checking the two headline figures โ 63 percent and 48 percent โ across coverage from three partner outlets, which reported matching numbers. On the strength of that three-way corroboration, our final determination on the claim is True.
