Washington, DC — On the day of the second annual AI Honors Gala, the Washington AI Network and Morning Consult released findings from a national poll of 1,501 U.S. adults examining how Americans use, trust, and want to govern artificial intelligence (AI). The survey, conducted May 27–30, 2026, comes as policymakers, industry, and the public face the rapid disruption and integration of AI into health, finance, work, and this year’s elections.
“Adoption is outrunning trust — and it’s already happening. In the past week alone, more than a quarter of Americans used AI for medical information, and just as many acted on an AI answer without checking it,” said Tammy Haddad, founder of the Washington AI Network. “Washington is debating the future of AI while the country is already living it — and one-third of Americans have never used an AI chatbot. That gap is the conversation Washington needs to be having.”
Key findings
- More than 70 million Americans turned to AI for medical information last week — and one in four acted on AI’s answer without checking it elsewhere. In the past seven days alone, 28% of U.S. adults, or more than one in four Americans, used an AI chatbot for information about a medical symptom or medication; 21% used it for money or taxes; and 21% used it for a legal question. Yet only 16% of Americans say they trust AI “a lot” to provide accurate information, highlighting a clear paradox between use and trust.
- Community concerns about AI data centers are real – but they are movable. Initial opposition runs 47% to 34% support. Yet when a full package of community benefits — subsidized internet, lower electric bills, 500 local jobs, school funding, renewable-energy guarantees, water-use protections — is on the table, support rises to 50% in favor with only 24% remaining opposed. No single benefit on its own offsets opposition. But a comprehensive slate of community benefits will.
- AI loses every head-to-head trust matchup, but the margins tell a more complicated story. When it comes to Americans’ trust of AI vs. another source, AI is most competitive against Congress (24% AI vs. 45% Congress) and large corporations (25% vs. 40%). It trails furthest behind doctors (16% vs. 63%) and close friends and family (16% vs. 61%). Among daily AI users, things shift: AI nearly ties Congress, beats local TV and gets close to corporations and tech CEOs. But trust relationships with AI are still forming – the “not sure” groups are 25-35% across each pairing.
- Most Americans don’t want AI replacing anyone — but when they do, they’ll sacrifice the suit before the scrubs. More than half (53%) don’t want AI to replace any profession asked about, but the pattern of who’s acceptable is telling. White-collar information brokers are the most expendable: customer service reps (17%), financial advisors (14%), Members of Congress (12%), and journalists (11%) top the list. Relational and skilled-trade roles are protected: hairdressers (5%), plumbers and electricians (5%), clergy (7%), and primary care doctors (8%) score lowest. The pattern suggests Americans are more open to AI in transactional or institutional roles than in relational ones.
- Congress holds a narrow edge in AI regulation — but 40% of Americans aren’t sure who they trust. When asked whose judgment they’d trust more on AI regulation — an AI company CEO or their member of Congress — Americans give Congress the edge, but not a mandate: 34% Congress, 26% CEO, with a striking 40% saying not sure. But among daily AI users, the CEO edges ahead: 44% vs. 34%. The public has not handed Congress a governing mandate on AI, even if it nominally prefers it.
- One in three voters plans to ask AI which candidate matches their views, but using AI on the campaign trail is a liability — unless your voters are already power users. Looking ahead to the November 2026 midterms, 44% of adults are likely to use an AI chatbot to fact-check a candidate, and 33% are likely to ask AI which candidate best matches their views. Finding out that a candidate used AI to write speeches, ads, or social media posts is a net negative: 41% say it would make them less likely to support that candidate, versus only 17% who say it would make them more likely. Among daily AI users, the dynamic nearly inverts: 43% say more likely. The authenticity penalty is real, but it’s concentrated outside the power-user cohort.
- 7 in 10 Adults Are Concerned About AI Having a Negative Effect on the 2026 Midterms. 70% of Americans, across party lines, are concerned about AI-generated deepfakes in the 2026 Election, and the same number are concerned about foreign government interference with AI and campaigns using AI to target voters personally.
- Americans are acting on AI answers they aren’t verifying. Only 15% of Americans always check AI answers against another source. Another 18% do so most of the time — meaning fewer than a third of users are consistently checking. 11% rarely verify, and 9% never do. Among daily users, the always-check rate rises to 34%, but that still means two-thirds of the heaviest AI users are not consistently verifying answers they’re acting on.
- On institutional trust, scientists are trusted. CEOs are not. Independent scientists are the most trusted to make decisions about AI (54% trust them ‘a lot’ or ‘some’). AI company CEOs are trusted ‘a lot’ by just 13%, with 34% saying ‘not at all.’ Congress comes in at 36%, state governments at 41%, and the general public at 47%. AI company CEOs are trusted “a lot” by only 13% of Americans, with 34% saying “not at all.”
America is split on AI — not by party or age, but by the heavy-user divide.
A central pattern across the survey is the gap between the 18% of Americans who use AI “several times a day” and everyone else. Heavy AI users trust AI company CEOs more than Congress (44% to 34%) — an inversion of the general public. They would be more likely, not less, to support a candidate who used AI (43% vs. 17%). Two-thirds support a local AI data center. And 58% acted on AI’s advice in the past week without checking it elsewhere. On nearly every question in the survey, the daily-user cohort answered 20 to 40 points apart from the country as a whole.
“What stands out in this data is how fast the behavioral story has overtaken the attitudinal one,” said Dr. Kyle Dropp, Co-Founder and President of Morning Consult. “Americans are using AI for the kinds of decisions — health, money, legal — where the stakes are highest, but trust hasn’t caught up to usage.”
Methodology and crosstabs
The Washington AI Network × Morning Consult survey was conducted online among 1,501 U.S. adults from May 27–30, 2026. Results were weighted to approximate a target sample of U.S. adults based on gender, age, race, educational attainment, region, gender by age, and race by educational attainment. The unweighted margin of error is ±2.5 percentage points at a 95% confidence interval.
Full topline results, charts, methodology, and demographic crosstabs are available to credentialed members of the media on request: media@washingtonainetwork.com.
About the AI Honors
The findings will be featured at the second annual AI Honors gala that evening, at the Waldorf Astoria in Washington, DC. The black-tie event convenes more than 400 leaders from government, industry, academia, and civil society to recognize those shaping the future of AI. 2026 sponsors include Americans for Prosperity, Anthropic, Amazon, Bloomberg, Booking Holdings, Build American AI, Capital Power, Cohere, Constellation Energy, Flex Association, General Catalyst Institute, General Motors, GlobalWIN, IBM, JPMorganChase, LinkedIn, Microsoft, NobleReach Foundation, NVIDIA, Public Private Strategies Institute, Qualcomm, Salesforce, Scale AI, Silverado Policy Accelerator, Special Competitive Studies Project, TikTok, and Uber.
Website: http://www.washingtonainetwork.com
###
About the Washington AI Network
Washington AI Network is a bipartisan forum that brings together a wide variety of stakeholders from government, civil society, the private sector, media, and academia to collaborate and foster meaningful conversations, partnerships, activations, and actionable initiatives that address the biggest opportunities and greatest challenges around AI.
About Morning Consult
Morning Consult is a global decision intelligence company changing how modern leaders make smarter, faster, better decisions. Pairing proprietary high-frequency data with applied artificial intelligence, Morning Consult delivers living intelligence that anticipates the future and illuminates the best way forward. Founded in 2014 and headquartered in New York with offices in Washington, DC, Chicago, and San Francisco, Morning Consult provides global survey research tools, data services, and intelligence to leaders in business, marketing, economics, and politics.