AI chatbot usage has surged to 21% of U.S. workers in 2025 — even as only about 17% of Americans believe the technology will benefit society over the next two decades, according to Pew Research Center. That’s roughly one in six people. Meanwhile, 57% of AI experts are optimistic about the same timeframe. The gap between the people building this technology and the people living with it has rarely looked this wide.

This isn’t a story about technophobes avoiding progress. It’s about a country that adopted a tool it fundamentally doesn’t trust — like leaving a one-star review for a restaurant you order from every week.

The Numbers Are Worse Than You Think

Job fears, existential dread, and a creativity crisis dominate public sentiment toward AI.

Stanford’s 2026 AI Index found that 64% of Americans expect AI to eliminate jobs over the next 20 years. Only 5% think it’ll create more. A YouGov poll puts it even starker: 77% of respondents consider AI a potential “risk to humanity.”

The specific fears, according to Pew Research Center, cut deep:

• 53% say AI will worsen creative thinking
• 50% believe it will damage people’s ability to form meaningful relationships
• 38% expect it to erode problem-solving skills
• The share of Americans “more concerned than excited” about AI climbed from 37% in 2021 to roughly half today

“While 57% of the public is highly concerned about AI leading to less connection between people, this drops to 37% among the experts we surveyed.” — Pew Research Center. And yet, usage keeps climbing. Among teens aged 13–17, 64% use AI chatbots, and about 30% use them daily, according to Pew. The generation most immersed in AI judges it most harshly — 61% of adults under 30 say AI will make people worse at thinking creatively, the highest rate of any age group.

Where Americans Actually Trust AI

Medical care earns cautious approval, but anything touching human connection gets a swift rejection.

Not everything registers as apocalyptic. Some 44% of Americans say AI will positively impact medical care over the next 20 years, versus just 19% expecting harm, according to Pew. Weather forecasting and data-heavy technical tasks attract similar goodwill. The dividing line is human contact: when AI functions as a precision instrument, people accept it. When it edges into relationships, creativity, or paychecks, trust collapses fast.

As a Newcomer analysis of 13 polls concluded, “economic disruption is the clear standout in what makes Americans the most nervous about AI.”

The oversight picture isn’t reassuring either. Americans split almost evenly on trusting the government to regulate AI — 44% trust, 47% distrust, per Pew. Majorities want stronger guardrails even at the cost of slower innovation. A notable gender divide persists: women view AI unfavorably by roughly 10 points, while men view it favorably by about 16, according to Data for Progress.

So here’s the structural irony worth watching: AI is winning on adoption while losing on legitimacy. That gap — between what Americans use daily and what they actually believe about its long-term impact — is precisely the pressure point that will shape AI regulation, workplace policy, and product design for the next decade. Whether that trust deficit narrows or widens may depend less on what the technology can do, and more on whether the companies and governments overseeing it can make a credible case that someone responsible is actually in charge.