Why do young adults use chatbots most yet trust them least?

Pew Research Center surveyed 5,119 U.S. adults (members of its American Trends Panel, representing the full U.S. adult population) from Feb. 17–23, 2026, and finds chatbot use "up substantially from the summer of 2024": about half of adults now report using AI chatbots, including roughly one-in-four daily. ChatGPT alone rose to 44% of adults, "up from 34% last year." But adoption does not track with favorable sentiment. Four-in-ten adults say AI will have a negative impact on society over the next 20 years, "far fewer believe its impact will be positive"; about two-thirds say AI is "advancing too fast" against just 2% who say too slowly; and roughly seven-in-ten predict AI will make their personal information less secure. Crucially, the heaviest users are the most negative: "adults under 50 are about twice as likely as those ages 50 and older to report using ChatGPT (57% vs. 28%)," yet "adults under age 30 are more likely to say AI will have a negative effect on society and on them personally than those 30 and older."

Pew offers no causal mechanism for the gap — the report is a cross-sectional snapshot, not a panel tracking the same respondents' attitudes over time. It does note that distrust of institutions compounds the picture: 67% of Americans have little to no confidence in the U.S. government to regulate AI effectively (up from 62% in 2024), and about six-in-ten are not confident in U.S. companies to "develop and use these tools responsibly." The report also flags a partisan realignment within that shared skepticism: Democrats' distrust of government AI regulation rose 20 points since 2024 while Republicans' fell from 70% to 61%, even as both parties' overall distrust levels remain high. Self-reported effects on productivity, creativity, happiness, and relationships are muted by comparison — three-in-ten say chatbots help their productivity and a similar share say they help keep them informed, with only 5% reporting harm on either measure, and most respondents say chatbots have no effect at all on happiness or relationships.

This complicates the practical-use framing in How does emotional chatbot use develop from practical use?: Pew's cross-sectional data confirms that practical tasks (search, work) dominate chatbot use nationally — only 13% use chatbots for news, one-in-ten for emotional support, and a smaller share for companionship — but shows that this practical, non-companionship-dominated usage pattern coexists with net-negative societal expectations rather than insulating users from them. It also sits in tension with the self-reported productivity gain in Does AI assistance help less experienced workers most?: that paper measured actual throughput in one firm's support workflow, while Pew's three-in-ten "helps my productivity" figure is self-report across the general population, a weaker form of evidence that this note is careful to keep distinct.

The survey cannot explain why heavier young-adult use coincides with lower optimism — whether through closer exposure to AI's failure modes, generational skepticism toward institutions generally, media diet, or some other factor Pew did not test — and a single cross-sectional wave cannot rule out that this is a stable generational difference rather than something usage itself produces. What it does establish is that, at the U.S. population level as of February 2026, rising adoption and rising skepticism are running in parallel rather than trading off against each other, which weighs against any simple assumption that familiarity with a technology breeds trust in it.