Poll Finds Voters Back Mandatory AI Safety Reviews but Also Favor Outright Ban and Distrust Government Oversight

A new poll shows an overwhelming majority of likely voters want powerful AI systems to undergo mandatory formal safety reviews before public release. The survey by the AI Policy Institute found that two-thirds of respondents preferred AI systems with guardrails over an outright ban when that option was presented.

When the choice was framed as no regulation versus a ban, voters strongly favored banning AI entirely.

Republicans were more enthusiastic than Democrats about government-led safety testing, though more than half of voters across parties supported the measure. Over 60 percent of both Republican and Democratic respondents said the federal government should set clear safety standards and evaluate companies' adherence.

More than 80 percent overall, including 84 percent of Democrats and 83 percent of Republicans, said AI companies should not build systems smarter than humans until they can demonstrate control.

Forty-seven percent of respondents said they would allow data centers if AI systems met safety and security standards, while 38 percent favored banning data centers entirely. Almost three-quarters of those surveyed said AI will become a more important political issue in the future.

The poll was released June 29, weeks after President Donald Trump signed an executive order in early June directing federal agencies to test new AI models for safety concerns on a voluntary basis.

On June 26, OpenAI released GPT-5.6 only to trusted partners after government safety requests, and the U.S. government cleared Anthropic to give trusted partners access to its Mythos 5 model. Peter Wildeford, director of policy at the AI Policy Network, said the results show Americans want stronger action.