The UK AI Compass is a large-scale segmentation of British public opinion on artificial intelligence. Drawing on a nationally representative survey of 2,911 UK adults (2,774 weighted), with oversamples of ethnic-minority and marginal-constituency voters, and five focus groups held across the nations and regions, it finds a public that has already formed strong, largely protective views on AI. Eighty-five per cent want stronger laws to make AI safe, and majorities favour regulation across most policy trade-offs. Rather than sorting people as simply excited or afraid, the report identifies six distinct AI mindsets that predict policy views, voting behaviour and message response more reliably than demographics. It maps how AI anxiety is reshaping electoral behaviour — in particular Labour's loss of voters to Reform and the Greens — tests five policy messages using real-time dial testing, and sets out a communication strategy for building a durable majority for sensible AI governance. Methods include latent class analysis, hierarchical value mapping, and real-time message dial testing.
Publication Date: 2026-06-05