Quinnipiac University’s March 2026 poll finds rising use of AI tools alongside persistently low trust in AI-generated information. Twenty-one per cent say they trust that information most or almost all of the time, a result the university describes as largely unchanged from April 2025.
The numbers: rising use, persistently low trust
The Quinnipiac survey paints a stark picture. A significant majority say they trust AI rarely or only sometimes. Meanwhile, most Americans are very or somewhat concerned about AI’s societal effects, and only a small percentage describe themselves as “very excited” about the technology. A majority now believe AI will do more harm than good in their daily lives, a significant increase from the previous year’s survey.
Quinnipiac computer science professor Chetan Jaiswal described a gap between growing adoption and limited trust.
The employment disconnect
The anxiety is sharpest around jobs. According to the poll, a significant majority of respondents believe AI advancements will reduce overall job opportunities — an increase from a year ago. Only a small percentage think AI will create more jobs. Yet the concern remains abstract for many: only a minority believe AI directly threatens their own position.
The fear has structural backing. Entry-level job postings in the U.S. have declined 35% since 2023, and Anthropic CEO Dario Amodei has publicly warned that AI could cause “unusually painful” disruption to the labour market.
Transparency and regulation gaps
The poll also finds concerns about transparency and regulation: a strong majority of respondents say businesses and government are not doing enough to be transparent about AI or to regulate it, respectively. Additionally, a majority oppose the construction of AI data centres in their communities, citing electricity costs and water usage.
Quinnipiac professor Tamilla Triantoro noted that greater familiarity with AI did not necessarily coincide with optimism about its employment effects.
What the gap reveals
The pattern captured in this data extends well beyond a consumer sentiment story. The poll shows that use and confidence need not move together: more Americans are using AI tools, while trust in AI-generated information has stayed low.
As Silicon Canals has previously explored, the underlying data infrastructure that powers these AI systems was already built before the public had any meaningful input into its governance. The Quinnipiac data suggests the public is now aware of this gap — and increasingly uneasy about it.
Jaiswal framed the findings as concern about uncertainty, trust, regulation and jobs rather than outright rejection of AI.
Feature image by Anna Tarazevich on Pexels
Correction — 4 October 2026: The earlier opening and subheading described trust as declining. Quinnipiac says it was largely unchanged from April 2025. The trend wording and the headline’s description of what the poll measured have been corrected.
