Public use of generative AI continues to grow rapidly, and while much of the public is undecided or distrusting, those who use AI systems tend to trust the tools they use and the information they provide. Despite growing use, familiarity, and considerable optimism among users, public attitudes towards the use of generative AI in news (and by politicians) are much more negative – and seem relatively settled despite much else changing.
First, weekly use continues to grow rapidly. Intentional use where people actively use generative AI systems (e.g. ChatGPT, Gemini, Copilot) has reached half the population in these high-income countries in less than four years, much faster than internet use, faster than smartphone use, and about as fast as social media use (at least going by US track record).
Second – in a context of wider public skepticism – those who use generative AI systems for informational purposes (answering factual questions, asking advice, getting news) generally say they trust the outputs.
Third, in parallel with intentional use, incidental generative AI exposure is even more widespread, as people, without necessarily wanting to, encountering AI-generated content in increasingly AI-infused legacy platforms and services, especially search engines, but also social media, as a feature on their devices, and on various other web sites (including news).
Fourth, even as there is a lot of change with intentional use growing and incidental exposure increasing, some public attitudes seem to be settling into a rather polarized pattern with the (growing number of) users far more optimistic than those who do not use (and in some cases fundamentally may not /want/ to use) generative AI systems.
These are some of the key findings from a new report by Felix Simon, myself, and Richard Fletcher and published by the Reuters Institute, based on surveys conducted in nine countries (Argentina, Australia, Denmark, Finland, France, Japan, Norway, the UK, and the USA) to explore how people use generative AI in their everyday lives, what they think its impact will be on different sectors of society, and what they think about its use in news and journalism specifically. The study builds on similar studies we did in 2024 and 2025.
