Bipartisan Poll: Overwhelming Majority of US Voters Back Mandatory AI Safety Reviews, With Republicans More Enthusiastic Than Democrats
Summary
- • Bipartisan majority of US likely voters supports mandatory safety reviews for AI before public release — beyond current voluntary Trump admin policy.
- • Republicans are now MORE supportive of government-led AI safety oversight than Democrats, reversing earlier polling trends.
- • Over 60% of both parties say the federal government — not AI companies — should set and enforce AI safety standards.
- • Voters prefer regulated AI over banning it, but would rather ban AI entirely than allow it with zero regulation.
Details
1,007 likely voters surveyed
Poll conducted June 10–11 by AI Policy Institute (AIPI), opt-in from an online research marketplace; represents bipartisan cross-section of likely US voters
Republicans more pro-regulation than Democrats
Significant reversal from earlier surveys where Republicans were more skeptical of government AI intervention; both parties now majority-supportive of mandatory reviews
Majority want mandatory pre-release safety reviews
Overwhelming majority of respondents support formal safety reviews before AI systems are released to the public — going further than current Trump admin voluntary opt-in policy
Federal government should set standards — not companies
Over 60% of both Republican and Democratic respondents want the federal government to establish and evaluate AI safety standards, not leave it to AI firms
47% back data centers with safety rules; 38% want full ban
Data centers have become a proxy for public fear of AI; voters prefer regulation to outright bans, but nearly 4 in 10 would ban them entirely
300+ data center bans and moratoriums nationwide
America's data center boom faces over 300 bans and moratoriums; independent researchers estimate blocked/delayed projects worth nearly $130 billion in 2026
Prefer regulated AI over banning; prefer banning over no regulation
When forced to choose between unregulated AI or outright ban, voters strongly prefer banning — revealing that the public draws a hard line against ungoverned AI systems
OpenAI GPT-5.6 limited; Anthropic Mythos 5 cleared
OpenAI was forced to limit GPT-5.6 to trusted partners due to government safety concerns; Anthropic's Mythos 5 was cleared for trusted-partner access — illustrating active government-industry friction the poll reflects
AI Policy Institute poll, June 2026 — US public opinion on AI regulation and safety oversight
What This Means
Public opinion on AI regulation has shifted meaningfully, with a bipartisan majority — including Republicans who previously resisted government intervention — now favoring mandatory safety reviews over the Trump administration's voluntary framework. This creates genuine legislative pressure: when both party bases want stronger rules, elected officials face risk by opposing them. The "regulate vs. ban" preference is particularly telling: Americans broadly accept AI, but only with meaningful guardrails — and would choose elimination over a lawless alternative. For AI developers, this signals that voluntary compliance postures may not hold politically, and regulatory mandates are increasingly likely.