Americans Grow More Wary of AI Even as 70% Say They Know the Tech Well

Americans know more about artificial intelligence than they did a year ago. And they like it less.

That’s the short version of a new Gallup report, and the numbers behind it are pretty striking. Seven in ten Americans now say they’re somewhat or extremely knowledgeable about AI. Yet 39% believe the technology does more harm than good — up from 31% in 2025. Only 9% say AI does more good than harm. The remaining 52% see it as a wash, roughly balanced between benefits and downsides. So familiarity isn’t breeding comfort. It’s breeding doubt.

The sharpest shift is happening among younger adults. Nearly half of Americans aged 18 to 29 now view AI as more harmful than helpful. A year ago, that figure sat at 36%. That’s a big jump in a short time, and it probably matters a lot to the tech industry, which has spent years pitching AI as the future that younger generations would naturally embrace.

Trust in Business AI Use Is Crumbling Fast

The trust numbers are worse. Only 27% of Americans say they trust businesses “somewhat” to use AI responsibly — down from 31% in 2025. Doesn’t sound like a huge drop, but look at the other end of the scale: the share of Americans expressing no trust at all rose from 29% to 41%. That’s not a drift. That’s a collapse.

Among adults aged 18 to 29, the numbers got even uglier. Trust in businesses using AI responsibly fell from 30% to 20% in a single year. So the same age group that’s growing more skeptical about AI’s impact is also the group that has lost the most faith in companies handling it well. For businesses betting heavily on AI-driven products and services, that’s a real problem.

Job fears are feeding into all of this. Nearly 79% of Americans now expect AI to reduce U.S. jobs over the next decade, up from 73% the prior year. Among younger adults, job loss concerns jumped from 62% to 75%. That’s a 13-point swing in twelve months. And it’s not an abstract worry — it’s the kind of number that shapes how people vote, what they buy, and whether they trust the companies building these systems.

Where Humans Still Win

Not everything in the report is doom for AI. Americans do see the technology as roughly on par with humans in certain areas — driving, financial advice, medical advice. That’s actually a meaningful shift from earlier years, when AI was seen as clearly inferior across the board.

But when the task requires judgment? Humans still win. Hiring decisions, creative work, helping students with schoolwork — respondents consistently rated human performance higher in every one of those categories. People are fine with AI driving a car or crunching numbers. They’re not fine with AI deciding who gets a job or grading an essay. That distinction is clear in the data, and it’s probably not going away soon.

Gallup’s findings aren’t happening in a vacuum. An NBC News poll from March found that 56% of Americans had used AI tools like ChatGPT, yet 57% saw more risks than benefits. That’s a majority of users who are worried about the very thing they’re using. And a June survey by Anthropic found job losses as a top concern across demographics and political affiliations — not just among workers in sectors typically seen as vulnerable.

The Anthropic survey also turned up a number worth sitting with: only 15% of Americans trust AI companies to make responsible decisions about the technology’s future. Fifteen percent. That’s a trust deficit so large it’s kind of hard to overstate. Businesses and policymakers can’t just assume the public will come around once the products get better. The skepticism seems to be running ahead of the technology itself.

A Widening Gap Between Use and Trust

Here’s the strange part. Usage keeps climbing. More Americans are engaging with AI tools in their daily lives even as their confidence in those tools — and the companies behind them — keeps falling. It’s a weird dynamic. People use it, they get more familiar with it, and they trust it less.

That gap between adoption and trust is probably the central challenge for anyone trying to build a sustainable AI business right now. You can get people to try the product. Getting them to believe in it is a different thing entirely.

The Gallup data puts hard numbers on something a lot of people in the industry have been reluctant to say out loud: the public relations problem around AI isn’t a messaging problem. It’s not that people don’t understand the technology. Seventy percent say they do. The problem is that the more they understand it, the less they like what they see.

Among the 18-to-29 cohort, trust in business AI use sits at just 20%.