New research from Verian shows that three quarters of adults (75%) are very/fairly concerned about AI being used irresponsibly. A similar proportion (74%) are very/fairly concerned about AI being used to spread misinformation. However, most adults (69%) are using AI tools, with one in five (21%) using these tools at least once a day.
Verian’s research took place between 12th and 15th December 2025, using our Public Voice random sample panel. Our restrictive method of panel-building provides benefits in terms of sample and data quality when compared to many other panels used in the UK for polling.
Most Britons are aware of AI tools, with nine in ten (90%) having heard of AI tools such as ChatGPT, Gemini, Co-Pilot, Claude, or Grok. Awareness is more or less universal (98%) among those aged under 35 and among those aged 65+,with three in four (77%) having heard of at least one of these tools.
Concerns about AI are widespread, with the following being the most selected concerns:
Despite widespread concerns surrounding AI, 69% of the population use it for at least one purpose. The most frequently cited uses:
Given a list of thirteen AI tool use cases (plus a ‘write-in’option), the average number of uses selected was 2 (mean=2.1) among all adults, rising to 3 (mean=3.1) among those using the tools. Younger users reported using a greater number of use cases than older users, with an average of 4 for 18-24 year olds (mean=3.9)and 25-34 year olds (mean=3.8) compared to 2 for 55-64 year olds (mean=2.3) and users aged 65+ (mean=1.8).
Overall, one in five (21%) adults use an AI tool of this type at least once a day, with peak daily use among those aged 25-34 (35%) and the lowest daily use among those aged 65 plus (8%).
One in five adults (21%) are using AI tools to get advice of some kind (legal, financial, or personal), including one in three (34%) 18-34 year olds. However, among those using these tools for this purpose, six in ten (60%) are concerned about the accuracy of AI generated results and three in four (75%) are concerned about the replacement of human with AI judgment in critical decisions.
A total of 1,427 interviews were conducted online among adults living in Great Britain between the afternoon of the 12th and morning of 15th December 2025. The survey data tables and further details on the methodological approach and weighting will be published on the Verian website by 18th December 2025.
Our sample was drawn from Verian’s random sample panel Public Voice. This panel is used extensively for social research commissioned by government, academic and third sector organisations, including those based in the US and Europe. Membership of this panel is restricted to those living in a controlled sample of UK addresses drawn from the Royal Mail’s master database. This restrictive method of panel-building is relatively expensive but will ordinarily provide benefits in terms of sample and data quality when compared with a panel that any adult resident in the UK can join. Although the short-period fieldwork web-only protocol used for this poll is much more limited than is typical for a social research survey, the demographic and political composition of the sample is only modestly degraded compared to what could be obtained using the full social research data collection protocol (two to three weeks using both web and telephone interview modes).
The data was weighted to match population totals for age, gender, 2024 General Election voting patterns, education, region, and likelihood to vote in the next General Election.
For further information on any of the data and findings from the report please contact PressEnquiries@veriangroup.com