UK Generative AI Use Hits 59%, but Trust Still Lags Adoption

Fifty-nine percent of UK adults used generative AI in the past three months, according to the first large-scale government count of the habit. Almost everyone has heard of it: 97% of adults reported at least some awareness of AI, the Department for Science, Innovation and Technology (DSIT) found.

Most of those users still do not fully trust what the technology tells them. Nearly three in four adults, 74%, said they worried AI-generated information could be inaccurate, even as they kept using it for work, study and everyday tasks.

UK Adults Cross the 59% Threshold on Generative AI

The figure comes from DSIT’s Public Engagement Survey, a continuous push-to-web study of UK adults aged 16 and over, delivered by the research firm Verian on the department’s behalf. Fieldwork ran from November 2025 to March 2026 and gathered 30,698 responses, one of the broadest official snapshots of AI use the UK has produced.

Under the headline number, the tools split roughly in half. Some 56% of adults had used generative tools that produce human-like text or speech, and 54% had used AI-powered digital assistants. The two categories overlap heavily rather than stacking on top of each other.

Ofcom, NatCen and DSIT Chart the Same Curve

DSIT’s number is not the first data point on this trend. It lands inside a run of tracking surveys that all point the same direction.

1. November 2024: NatCen surveyed the public on behalf of the Ada Lovelace Institute and the Alan Turing Institute, building on a 2022 baseline to track how attitudes toward AI applications were shifting.
2. 2024: Ofcom’s own separate tracker put UK adult use of chatbot-style AI tools at 31%.
3. 2025: The same Ofcom tracker recorded that figure climbing to 54%, with ChatGPT visits from UK users rising sharply over the same period.
4. November 2025 to March 2026: DSIT’s Public Engagement Survey runs its fieldwork through Verian, collecting 30,698 responses nationwide.
5. July 2026: DSIT publishes its first results, putting overall adult adoption of generative AI at 59%.

Ofcom and DSIT ask different questions and measure different things, so their numbers should not be read as the same metric. But NatCen’s own research for the Ada Lovelace and Alan Turing Institutes shows the underlying pattern has been building for years, not months.

London Graduates Pull Away From Everyone Else

Adoption was highest among adults aged 16 to 44 and fell steadily after that, reaching just 8% among people aged 85 and over. Education widened the gap further: only 19% of adults with no formal qualifications had used generative AI, against 75% of degree holders, nearly four times as many.

Geography added another layer. London recorded the highest regional adoption rate at 69%, 18 percentage points above North East England’s 51%.

The split echoes a pattern already showing up one rung down the education ladder. Among students heading into the workforce, nearly all now use AI tools daily, though just 28% were formally taught how to use them, a gap that could follow this cohort straight into their careers.

Time, Not Curiosity, Drives the Habit

Among online respondents who use generative AI, 83% reported personal use, 54% used it for work, and 27% for education or study. Saving time was the leading motivation, ahead of generating ideas and summarising information. Easier access to information or advice was rated the technology’s biggest benefit.

• Saving time was the single most common reason adults gave for using generative AI.
• Generating ideas came second, spanning both personal and work contexts.
• Summarising information rounded out the top three, tied closely to time savings.

That workplace pull shows up in separate industry research too. A 2025 global study from the University of Melbourne and KPMG found almost two-thirds of UK workers, 65%, say they intentionally use AI at work, with 44% worried about being left behind if they don’t. The individual habit is running well ahead of any organisational payoff. Across UK businesses more broadly, just 6% report a real financial return on their AI adoption, even as adoption itself becomes close to universal.

Why Do Most Users Still Not Trust the Answer?

Most UK adults now use generative AI regularly, but a majority still don’t fully trust what it produces. DSIT found 74% worried AI-generated information could be wrong, and separate employer and public-opinion research helps explain why that doubt persists even as use keeps climbing.

Workplace transparency plays a part. Deloitte’s research on UK businesses found 58% of UK employees using GenAI tools agree their organisation is transparent about the technology’s impact on their role, a figure that shapes how comfortable staff feel relying on the output.

Regulation itself seems to ease some of that unease. According to the Ada Lovelace and Alan Turing Institutes’ own public attitudes tracker, nearly three-quarters of Britons now say laws and regulation would make them more comfortable with AI, up from 62% in 2023. DSIT’s survey found the same appetite: at least two-thirds of respondents backed legislation to improve AI safety, increase transparency, reduce bias and limit job displacement.

Westminster Has a Bill It Won’t Write

That demand for legislation runs into a stalled process. As of mid-2026, no single AI statute has cleared Parliament. A House of Lords debate on an AI Regulation Bill was still working through committee stage in June 2026, with peers arguing that legislation to protect public safety and transparency was long overdue.

The Ada Lovelace Institute, a research body focused on AI and data, has been blunt about the gap between ambition and safeguards.

Since coming into power, the current UK government has shown unchecked optimism about AI’s potential, shifting its focus from AI safety to AI adoption.

The institute noted the government has committed £2 billion to expanding AI by 2030 without setting out matching rules on disclosure, leaving regulators unable to weigh the risks its own International AI Safety Report had flagged.

DSIT has said official statistics from the Public Engagement Survey will be published annually. The 59% figure logged this July is the opening entry in what is meant to become a running national record, the first of many checks on whether trust in AI ever catches up with how many people are already using it.