Exclusive: Americans are using and worrying about AI more than ever, survey finds
Add Axios as your preferred source to
see more of our stories on Google.

Illustration: Shoshana Gordon/Axios
Nearly seven in 10 Americans have used AI, but concerns around the technology's impact on jobs and education is on the rise, per a new report from AI governance nonprofit Fathom.
Why it matters: People in the U.S. are steadily using AI more, but many say they aren't convinced that it's helping them and that rules would make the tech more trustworthy.
Driving the news: When asked who they trust to "ensure that AI is developed with proper guardrails," federal government agencies came in last, with just 42% of Americans expressing confidence in their ability to regulate AI.
- 55% of Americans surveyed said they trust "independent subject matter experts" the most, followed by tech companies at 50%, per the report seen first by Axios.
By the numbers: 82% of those polled said they had heard of or seen AI, up from 77% in 2024.
- Of the nearly seven in 10 people polled who said they had used AI, nearly half reported using it daily or weekly.
- Concern over the technology has increased to 33% from 30% the previous year.
Nearly half of respondents, or 49%, said they think that AI will create "significant job loss" and disrupt the labor market more than previous technologies.
- More than 75% of those polled by Fathom said there should be rules around keeping children safe online when they use AI.
People say they're primarily using AI for general information inquiries, through search engines and social media rather than standalone tools.
Methodology: Fathom's annual report is based on a national survey of 2,036 interviews conducted in August and September and has a margin of sampling error of plus or minus 2.17 percentage points at the 95% confidence level.
- The group also held four virtual focus groups among Ohio and California voters in August, featuring Ohio Republicans, Ohio independents, California Democrats, and California Republicans.
Context: Fathom works on governance models for state legislatures, including one on independent verification organizations that Axios wrote about earlier this year.
- That idea ended up in a California bill and is expected to pop up in other states next year.
What they're saying: "Americans understand that the pace of AI innovation is too rapid for traditional models of government regulation to keep pace, but they also don't trust AI companies alone to prioritize safety and security over profits," Andrew Freedman, chief strategy officer and co-founder of Fathom.
- "Our latest research shows Americans want a voice in how AI develops, with strong bipartisan agreement on the need for guardrails. This creates space for new approaches to AI oversight."
