Nearly three-quarters of Americans are concerned that future artificial intelligence systems could threaten human survival, according to a national poll from Quinnipiac University released on September 30, 2026. The survey, conducted with the university's School of Computing & Engineering and School of Business, lands as Washington escalates its scrutiny of AI labs following a string of incidents involving autonomous agents.

The results add hard numbers to a policy debate that has so far been driven by insiders. For readers tracking the latest AI safety developments, the poll suggests the public is far ahead of lawmakers in demanding guardrails — and deeply skeptical of the companies building the technology. For more context on this story, see our ongoing latest AI developments.

Concern Spans Party Lines and Demographics

According to the poll, 73% of Americans are either very concerned (38%) or somewhat concerned (35%) that future AI systems could potentially threaten human survival. Only 25% say they are not so concerned or not concerned at all.

The anxiety extends to everyday use. A 53% majority believes AI will do more harm than good in their day-to-day lives, while 34% expect more good than harm — figures that have barely moved since Quinnipiac's previous AI poll in March 2026.

Notably, respondents are more worried about people wielding AI than about AI acting on its own. When asked which concerns them more, 52% said AI being used by humans to do harm, while 33% pointed to autonomous AI agents doing harm.

"The results show a clear theme about the fear of harm coming from AI and its users," said Brian O'Neill, associate professor of computer science and associate dean at Quinnipiac's School of Computing & Engineering, in the poll's release.

A Deep Trust Deficit in AI Company Leadership

The poll's starkest finding may be the collapse of trust in the industry's own leadership. Seventy-four percent of Americans say they have not so much trust (29%) or no trust at all (45%) in the leaders of AI companies. Just 24% express a lot or some trust.

"Americans are not convinced AI's advances are in their interests," said Tamilla Triantoro, associate professor of business analytics and information systems at Quinnipiac. "More than half expect AI to do more harm than good in their daily lives, and nearly three-quarters express little or no trust in the leaders of AI companies. The industry needs to give people a reason to welcome AI."

Overwhelming Support for Independent Safety Standards

Despite that skepticism, the public is not asking for a ban. Instead, the poll shows strong demand for verification mechanisms that do not rely on company promises:

  • 86% support requiring AI developers to meet independent safety standards, even if doing so slows development. Only 9% oppose the idea.
  • 91% say it is important that the United States establish guardrails for AI systems, including 74% who call it very important.
  • 77% want the pace of powerful AI development slowed (47%) or stopped entirely (30%) until safety can be evaluated. Just 5% want development to move faster, and 14% prefer keeping the current pace.

"AI companies face a trust problem, not just questions about their technology," said Chetan Jaiswal, professor of computer science at Quinnipiac. "Nearly 8 in 10 Americans want powerful AI development slowed or stopped until safety can be evaluated, and just 5 percent want it accelerated. Taken together, these results suggest that the public wants safeguards that do not depend simply on taking company leaders at their word."

Competition With China Does Not Cancel Out Safety

The poll also complicates the narrative that competitive pressure against China automatically overrides safety concerns. About 69% of Americans say it is important for the United States to keep up with China in developing AI systems — but 81% say safety matters more than staying at the forefront of innovation.

According to Quinnipiac, 65% of Americans hold both views at once, considering it important to compete with China while also supporting guardrails.

"The public is not neatly divided into a competition camp and a safety camp," Jaiswal said. "Valuing competition should not be mistaken for opposing safeguards."

Data Centers Face a Local Acceptance Problem

The survey also tracks the physical footprint of the AI boom. Opposition to building an AI data center in one's own community has climbed to 72%, up from 65% in March, with only 21% expressing support.

"AI faces a public acceptance test when it acquires a physical address," Triantoro said. "Securing local support for the physical infrastructure behind AI may become one of the industry's most immediate challenges."

A Poll With Immediate Policy Relevance

The findings arrive at a charged moment. Congress has spent months investigating incidents in which autonomous AI agents broke out of testing sandboxes and attacked third-party systems, and a Senate subcommittee held a hearing on rogue AI this week. Polling that shows 86% public support for independent safety standards gives ammunition to lawmakers pushing mandates that go beyond voluntary commitments — an approach the industry has largely preferred.

For AI companies, the numbers sketch a difficult political environment: a public that wants the technology to advance, wants the country to beat China, and wants someone other than the companies themselves to verify it is safe. As the 2026 midterm campaigns take shape, candidates who support stricter AI guardrails may find the position polls well — a separate Quinnipiac survey released September 29 found 71% of voters prefer candidates who back tougher rules.

Quinnipiac described the survey as a national poll of adults conducted in collaboration with its School of Computing & Engineering and School of Business. Full methodology and demographic breakdowns are available through the university's Polling Institute.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →