Key points
- YouGov data: 89% of U.S. adults have at least one AI concern, yet 55% use AI tools weekly, including 55% of those with concerns.
- Top worries: data privacy and security at 63%, over-reliance at 53%, distrust of AI companies at 48%, accuracy at 46%.
- Weekly users and non-users sit 46 points apart on AI's societal impact: 52% positive versus 6%.
- Pew (June 2026) found 52% more concerned than excited, a figure that has stayed between 50% and 52% since 2023.
Why it matters for marketers: Heavy AI use does not signal trust, so marketers should segment research by usage frequency and answer privacy and accuracy worries directly instead of assuming adoption equals approval.
Fresh survey data points to a gap between how much people use AI and how much they trust it. In a YouGov report by Rishad Dsouza, 89% of U.S. adults named at least one concern about AI, while 55% said they use AI tools weekly. People with concerns used AI at nearly the same rate as people with none, 55% versus 61%.
Pew Research Center's June 2026 survey adds context: 52% of adults are more concerned than excited, 37% feel both equally, and 9% are more excited. Search Engine Journal contributor Greg Jarboe notes that this reads as ambivalence more than hostility, and that the concern figure has barely moved since 2023.
For marketers, the useful split is by usage frequency. Among weekly users, 52% say AI's impact on society has been positive; among people who never use it, 6% do. YouGov did not ask whether people trust what AI says about a brand, so Jarboe treats the figures as a segmentation signal, not a trust score.
Jarboe's suggestions are practical. Add a usage-frequency question to surveys and cut results by it. Write separate pages and ads for heavy users and non-users. Address privacy and accuracy worries by citing primary sources, naming authors, dating data and explaining how user data is handled.
He also suggests testing a landing page that states a product's limits against one that promises everything, measuring qualified leads and sales conversations rather than clicks. He frames the plainer page winning as a hypothesis, not a proven result, so teams should run the test before changing messaging broadly.
Sources
This brief is our own summary. Read the full reporting at the source.
