Reuters Institute finds over half of surveyed users see AI-generated search answers weekly, trust remains moderate

The Reuters Institute for the Study of Journalism published its 2025 Digital News Report, an annual survey run across dozens of countries, with a section devoted for the first time to how audiences encounter and trust AI-generated content in search and news. Across the surveyed markets, 54% of respondents said they had seen an AI-generated answer to a search in the past week — more than the 34% who reported using any generative AI tool directly in the same period, and higher still in some countries: 70% in Argentina, 64% in the UK, 61% in the US, against a low of 29% in France.

Trust in those answers was mixed. Of the people who had seen an AI-generated search result, roughly half said they trusted it. Across the full sample, including people who had not encountered such answers, trust in AI-derived answers ran well below trust in news generally — the report found respondents expected AI to make news less transparent, less accurate and markedly less trustworthy than human-produced journalism, even as they used AI search tools for convenience.

Google’s AI Overviews specifically showed a lower weekly reach (4%) than the aggregate figure for AI-generated search answers, which the report’s authors called surprising given the feature’s prominence at the top of results pages — a gap they attributed in part to some respondents not recognising AI Overviews as AI-generated. The report’s authors linked persistent trust gaps partly to well-publicised cases of generative AI producing factual errors. As one of the largest recurring cross-national surveys of news consumption, the finding offered an early, if imprecise, measure of how quickly AI-mediated search was displacing traditional links to publishers, and how little audiences yet trusted what replaced them.