AI Trust Gap: Daily Users Still Worry About AI Risks

What happened

A newly published report finds that Americans who use artificial intelligence tools on a daily basis are just as concerned about the technology's risks as those who rarely or never use it. The research also finds comparable levels of support for AI regulation across both groups, suggesting that familiarity with AI tools does not translate into greater reassurance about their safety or governance.

The finding challenges a common assumption in technology adoption — that regular, hands-on use of a product naturally builds trust and reduces anxiety about it. According to the report, that pattern does not hold for AI: frequent users remain wary of issues such as misuse, job disruption, misinformation and lack of oversight, even as they continue to integrate the tools into their daily routines.

Why it matters

For organisations rolling out AI-powered products, services or internal tools, this is a signal that usage and trust are separate variables that must be managed independently. Leaders have often treated adoption metrics — logins, query volume, feature uptake — as a proxy for confidence. This report suggests that people can use a technology heavily while still harbouring real concern about its risks, which has direct implications for how AI features are communicated, disclosed and governed inside customer- and employee-facing products.

It also reinforces that regulatory appetite is not a niche position held only by sceptical non-users. Even the most engaged users of AI tools appear to want guardrails, which strengthens the case for organisations to get ahead of compliance and transparency expectations rather than assuming heavy usage equates to public comfort with lighter-touch oversight.

The Renascence take

Most coverage of AI adoption still conflates usage with acceptance — if people are using the tool, the thinking goes, they must be comfortable with it. This report quietly dismantles that assumption, and it has a direct parallel in service design: customers can tolerate, rely on and even depend on an experience while still distrusting the system behind it.

The behavioral lesson here is that trust is not a byproduct of repetition — it has to be designed for deliberately, through visible controls, honest disclosure and clear escalation paths when AI gets something wrong. Any organisation reading strong AI usage numbers as a mandate to relax on transparency or governance is misreading its own customers. The operators who win long-term will treat "high usage, high anxiety" as the real baseline, and build reassurance into the experience itself rather than assuming adoption will eventually do that work on its own.