The 5 Americas of AI Concern
How do Americans really feel about artificial intelligence, and how do they respond to different arguments about its risks?
Each dot represents a survey respondent. Together, they form the shape of a nation grappling with a technology that could reshape everything.
n = 2,735 · Scroll to explore ↓
Concern about AI is rising
But why, and among whom?
Artificial intelligence is transforming how we work, communicate, and make decisions. Public opinion is shifting fast.
Each dot represents a respondent in our nationally representative sample. Right now they are spread across the country, undifferentiated. But beneath the surface, Americans hold very different views about what AI means for them.
Concern is rising
From minority view to mainstream worry
In 2021, Pew Research found that 37% of Americans said they felt more concerned than excited about AI in daily life. Concern was real, but still a minority position.
Then ChatGPT launched. By mid-2023, that number had jumped to 52%, a 15-point surge in less than two years. AI was no longer abstract.
By 2025, concern had plateaued around 50%. Half of Americans remained more worried than excited, even as AI became embedded in daily tools.
Our survey asked a broader question: whether respondents were at least somewhat concerned about AI as a national issue. We found 73%. A single number that hides enormous variation.
Our sample
Who did we talk to?
We recruited a nationally representative sample of 2,735 U.S. adults through Prolific. 73% expressed at least some concern about AI and completed the full survey, forming our analytical sample of 2,000.
Zooming in
The concerned public is not a monolith
Let the unconcerned dots fade away. We are left with the 73% who said they worry about AI in some form.
Using latent class analysis, we identified five distinct segments within the concerned public. Each one has a different profile of fears, beliefs, and trust levels.
The Five Americas of AI Concern
Five distinct ways of worrying about AI
The circle splits. Five groups emerge, each colored by their dominant orientation toward AI risk. The largest accounts for nearly one in four concerned Americans.
Click any segment to explore it in depth, or keep scrolling for a guided tour of all five groups.
The Concern Landscape
What types of AI harm worry each group most?
Americans worry about many different aspects of AI. We asked respondents to pick their top 3 concerns from 14 options. Each bubble represents a unique combination of three concerns, positioned near the concerns it contains. Bigger bubbles mean more people chose that exact trio.
Which Messages Resonate?
Nine policy framings, tested across all five segments
We tested nine different policy messages about AI with our respondents. Each column shows how one segment rated the message: green squares mean 'very convincing,' red means 'not at all.'
Explore Each Segment
Click on a segment below to dive deeper into its demographics, concerns, and policy attitudes, and responses to different framings of AI risk.