Back to AI & Tech

Americans want better AI safety guardrails

Source reportMethodology

Overview

Support for AI safety regulation is broad and lopsided. About nine in ten U.S. adults back mandatory disclosure of safety-test results (89%) and independent review of the most powerful systems before deployment (89%), and 80% support federal authority to block a risky system.


The public also doubts the industry's own proposals. A 43% plurality say industry-drafted regulations are designed mainly to benefit the companies, while just 9% say they are mainly designed to protect the public.

Topline

89% of U.S. adults support mandatory safety-test disclosure, 89% support independent pre-deployment review and 80% support federal authority to block a risky AI system.

Each of the following AI regulations? - Requiring AI companies to publicly disclose the results of safety testing for their most powerful AI systems

  • Strongly support 54.8%
  • Support 34.4%
  • Oppose 7.9%
  • Strongly oppose 2.9%

2026 · base n 1,690 · +/- 2.8%

AI Regulation

View source data

Disclosure and independent review draw near-consensus support

89% of U.S. adults support requiring AI companies to publicly disclose the results of safety testing for their most powerful systems (55% strongly support plus 34% support); 11% oppose (8% oppose plus 3% strongly oppose).

Requiring independent, third-party review of the most powerful systems before deployment draws the same breadth: 89% support it (53% strongly plus 36%), while 11% are opposed.

Older adults are the most emphatic. On disclosure, 63% of adults 65 and older strongly support the requirement, compared with 47% of adults under 30.

Most would let the government stop a risky system

Four in five adults (80%, combining 45% strong support and 36% support) favor giving the federal government the authority to block the release of an AI system found to pose a risk. Just 6% strongly oppose it.

When the government and AI companies disagree over safety, 49% say the government should generally have the final say, about three times the 16% who would leave the call to the companies. Another 35% are not sure.

Topline

43% of U.S. adults say the AI industry's proposed regulations are designed mainly to benefit the companies; 9% say they mainly protect the public.

Thinking about these proposals, which of the following best reflects your view?

  • These proposals are primarily designed to benefit the companies — by shaping regulations in ways that favor establishe 43.3%
  • I'm not sure / I don't have enough information to say 32.8%
  • These proposals are equally motivated by public interest and company self-interest 14.8%
  • These proposals are primarily designed to protect the public — the companies are genuinely trying to prevent harm from 9.1%

2026 · base n 1,690 · +/- 2.8%

AI Regulation

View source data

The industry's own proposals meet suspicion

Asked about the regulations AI companies have proposed for their own industry, a 43% plurality say the proposals are primarily designed to benefit the companies, 15% say they are equally motivated by public interest and self-interest, and just 9% say they are primarily designed to protect the public. Another 33% say they do not have enough information to judge.

Asked who they trust more to determine whether a model is safe, 56% of adults say they are not familiar enough with the situation to have an opinion. Among the rest, about twice as many trust the U.S. government (30%) as trust the AI company, Anthropic (14%).

Methodology

Full methodology
Mode
Verasight panel recruited via random address-based sampling, random person-to-person text messaging, and dynamic online targeting
Population
US adults age 18+
Field dates
2026-06-18 → 2026-06-19
Base (unweighted)
1,690
Margin of error
+/- 2.8%
Module
AI Regulation
Sponsor
Verasight
Weight variable
weight
Weighting targets
age, race/ethnicity, sex, income, education, region, metropolitan status

Sources

[7]
  • 01
    To what extent do you support or oppose each of the following AI regulations? - Requiring AI companies to publicly disclose the results of safety testing for their most powerful AI systemsShows support for requiring public disclosure of AI safety-test results.reports.verasight.io/reports/june-2026-ai-survey
  • 02
    To what extent do you support or oppose each of the following AI regulations? - Requiring AI companies to have their most powerful AI systems reviewed by independent, third-party evaluators before deploymentShows support for independent third-party review before deployment.reports.verasight.io/reports/june-2026-ai-survey
  • 03
    To what extent do you support or oppose each of the following AI regulations? - giving the federal government the authority to block the release of an AI system if it is found to pose a riskShows support for federal authority to block risky AI systems.reports.verasight.io/reports/june-2026-ai-survey
  • 04
    Thinking about these proposals, which of the following best reflects your view?Shows how adults read the motives behind industry-proposed regulations.reports.verasight.io/reports/june-2026-ai-survey
  • 05
    Which of the following best reflects your general view on AI safety disputes between the government and AI companies?Shows who adults think should have the final say in AI safety disputes.reports.verasight.io/reports/june-2026-ai-survey
  • 06
    If you had to pick, who do you trust more to determine whether the model is safe?Adds who adults trust more to judge whether a model is safe.reports.verasight.io/reports/june-2026-ai-survey
  • 07
    Do you support or oppose this proposal?Adds support for a public-ownership proposal for AI companies.reports.verasight.io/reports/june-2026-ai-survey

Citation

What do Americans from both parties agree on? AI Regulation, fielded June 18-19, 2026, N=1,690 US adults age 18+, +/- 2.8%.

https://reports.verasight.io/reports/june-2026-ai-survey#to-what-extent-do-you-support-or-oppose-each-of-the-following-ai-regulations-requiring-ai-companies-to-publicly-disclose-the-results-of-safety-testing-for-their-most-powerful-ai-systems

Verasight survey methodology

How Verasight conducts surveys.

This page describes the Verasight general survey contract, separate from how the Data Library packages it. Each wave's specific field dates, sample sizes, and module breakdown are listed in that wave's report.

Mode
Verasight panel recruited via random address-based sampling, random person-to-person text messaging, and dynamic online targeting.
Population
US adults age 18+.
Sample design
Surveys are run as omnibus or single-topic waves. Omnibus waves are split into modules with their own respondent set, typically around one thousand respondents per module.
Field window
Each wave specifies its own field dates. Most omnibus waves field across roughly two weeks.
Weighting
Per-module weighting to CPS targets including age, race and ethnicity, sex, income, education, region, and metropolitan status.
Partisanship benchmark
Pew Research Center's NPORS benchmarking surveys, three-year running average.
Vote benchmark
2024 presidential vote population benchmarks.
Margin of error
Typically about plus or minus 3.4 to 3.6 percent per module at standard module sizes. Question-level MoE is recomputed when a base shrinks materially below the module baseline.
Reporting
Every wave is published as a standalone report at verasight.io/reports with full instrument and methodology.
Transparency
Verasight is a member of the American Association for Public Opinion Research Transparency Initiative.

Wave-specific methodology, full weighting variable lists, and verbatim instrument text live in each report at verasight.io/reports.