Seventy-seven percent of US adults want AI development slowed or stopped until safety is verified, according to a Quinnipiac University poll. Forty-seven percent favored reducing the pace and 30% wanted a full pause, while 5% supported faster development.
Independent safety standards drew support from 86% of respondents even if they delay progress. The survey also found that 74% had little or no trust in AI-company leaders, 72% opposed an AI data center in their own community and 91% considered guardrails for US AI systems important.
Concern did not eliminate geopolitical pressure. Sixty-nine percent said keeping pace with China matters. Respondents were also more worried about people misusing AI than about autonomous systems by a narrower margin: 52% selected human misuse. Separately, 73% worried that AI could eventually threaten humanity’s survival.
Quinnipiac surveyed 1,202 US adults from September 24 through 27 and reported a margin of error of plus or minus 3.5 percentage points. As with any poll, wording and sampling affect the result, and stated preferences do not automatically predict voting or product use. The findings nevertheless show broad demand for safeguards that do not rely only on companies assessing their own systems.