The key UK statistics on AI and personal data in one place — how the public feels about AI, how many adults use chatbots, how many businesses have adopted AI and whether they govern it, and what people really think about machines making decisions about them, with the data period stated next to every figure.

This page combines the ICO’s March 2026 public survey, the UK Business Data Survey 2026, the latest ONS business AI adoption release and dated research from DSIT, Ofcom and the Ada Lovelace and Alan Turing institutes. Each uses its own questions and population, so the results should not be treated as one continuous time series.

It is scoped to attitudes, adoption and governance. Breach and enforcement figures live on separate pages — see our data breach statistics and GDPR fines statistics guides for those.

Key facts and figures

  • 76% of UK adults are concerned about computers making decisions about people, including 41% very concerned (March 2026).
  • 93% say it is important that AI-generated content is clearly labelled — 80% say very important (March 2026).
  • 30% of UK businesses reported using at least one AI technology in September 2026; the figure was 51% among businesses with 250 or more employees (ONS).
  • 41% of UK businesses that handle digitised data use AI, rising to 82% of large businesses (October 2025 – January 2026).
  • 17% of AI-using UK businesses have any policy or guidelines on AI use — just 5% have a formal written policy.
  • 60% reported chatbot use in the previous three months in DSIT’s July–August 2024 online survey; monthly personal use was 40% and work use 28%.
  • 72% of the UK public say laws and regulation would increase their comfort with AI, up from 62% in 2022/23 (November 2024).
  • 1.8 billion UK visits to ChatGPT in the first eight months of 2025 — up from 368 million a year earlier.

Checked on 15 October 2026. The newest business adoption figures are from ONS on 8 October 2026. Other figures retain their survey dates. Chatbot, generative-AI and broader AI-use questions have different definitions and should not be directly compared.

What does the UK public think about AI and their data?

DSIT’s Wave 4 online survey, conducted in July–August 2024, found 96% awareness of AI. Its executive summary describes 4,947 online interviews and reports a separate digitally disengaged study. Most online respondents could give at least a partial explanation, but awareness did not imply detailed understanding.

In the same 2024 online survey, 43% expected AI to have a positive societal impact and 33% expected a negative one. These are dated attitudes and should not be presented as the current balance of opinion without a comparable new wave.

DSIT’s Wave 4 trust figures were 85% for the NHS, 76% for academic researchers, 38% for government and 33% for social media companies. The comparison concerns trust in those organisations’ data use, not trust in every AI system they might deploy.

Public acceptance varies with the organisation and proposed use. Attitude surveys can help identify concerns, but do not establish whether an AI system is lawful, accurate or safe.

How many UK adults use AI chatbots?

DSIT’s 2024 tracker found six in ten online respondents had used an AI chatbot in the previous three months. Forty per cent reported monthly personal use and 28% monthly work use; these groups overlap. The finding is a dated measure, not a current adoption estimate.

The ICO’s March 2026 YouGov survey found 74% of UK adults had ever interacted with an AI chatbot. Its question included customer-service bots and virtual assistants such as Alexa, so it is broader than use of generative chatbots alone. Fifty-three per cent selected disclosure that a chatbot is AI as one of their preferred rules.

The Ada Lovelace and Alan Turing institutes’ 2024/25 research found 61% had heard of large language models and 40% had used one. Different terminology and questions mean this cannot be used to calculate how many respondents misunderstand the technology behind their chatbot.

Ofcom’s December 2025 Online Nation summary reported 1.8 billion UK ChatGPT visits in January–August 2025, versus 368 million in the same months of 2024. Visits are not unique users. It also reported that around 30% of Google searches displayed AI overviews at the time of its research.

How many UK businesses use AI?

ONS reported that 30% of UK businesses used at least one AI technology in September 2026, rising to 51% among businesses with 250 or more employees. The bulletin was published on 8 October from Wave 165, collected from 21 September to 4 October. BICS is a voluntary survey producing weighted estimates, designated official statistics in development; it is not a census of every business.

DSIT's UK Business Data Survey 2026 (fieldwork October 2025 – January 2026, n=4,450, published 18 June 2026) measures a narrower base — businesses that handle digitised data — and finds 41% of them using AI for at least one purpose. The size gradient is steep:

Business sizeUsing AIData period
Sole traders40%Oct 2025 – Jan 2026
Micro (1–9 employees)41%Oct 2025 – Jan 2026
Small (10–49 employees)51%Oct 2025 – Jan 2026
Medium (50–249 employees)58%Oct 2025 – Jan 2026
Large (250+ employees)82%Oct 2025 – Jan 2026
All businesses handling digitised data41%Oct 2025 – Jan 2026

The ONS and UK Business Data Survey estimates have different coverage, questions and reference periods. UKBDS uses businesses handling digitised data and explicitly includes free, paid and in-house tools. The difference between 30% and 41% is not a measured increase or contradiction.

Among businesses handling digitised data, UKBDS found 28% used AI to research information and 21% to summarise or collect information or draft correspondence. Among the narrower group using AI, 5% reported automated decision-making tools. These denominators differ, and the survey’s operational definition of automated decisions should not be substituted for the legal test of a significant solely automated decision.

Do UK businesses have AI policies and governance?

UKBDS chapter 3 reports that 17% of AI-using businesses had a policy or guidance on AI use and development: 5% formal and 12% informal. Having a policy alone does not demonstrate compliance.

Among AI-using businesses, 56% of large businesses reported a formal written policy, compared with 3% of sole traders. These are self-reported arrangements, not an audit of the quality of governance or proof that a particular business lacks controls.

Fifty-three per cent of AI-using businesses had at least heard of relevant regulatory guidance. Among those aware of the guidance, 19% agreed it was clear and easy to understand. The second percentage uses the aware subgroup, not all AI-using businesses.

The same survey found that 21% of AI-using businesses had integrated AI tools into existing business systems. Integration can change how information is accessed and shared, so governance should cover actual use, permissions and data flows. Our cyber skills guide covers the separate workforce survey.

Are UK adults comfortable with automated decisions?

The ICO’s March 2026 survey found 76% of UK adults concerned to some degree about computers making decisions about people, including 41% very concerned. YouGov interviewed 2,157 adults online on 26–27 March and weighted the results to UK adults aged 18 and over.

In questions about AI use, 52% found fraud detection acceptable, compared with 15% for hiring and 14% for benefits decisions. A separate question explicitly excluding human involvement produced different results, including 19% acceptance for recruitment shortlisting and 16% for benefit eligibility. AI assistance and solely automated decisions are different measures.

Confidence in pushing back is finely balanced: 44% of UK adults would not feel confident challenging a computer-made decision they believed was wrong, while 46% would (March 2026). And awareness of where automated decisions already happen is thin — only 18% of the public knew AI is used in welfare benefits assessments (Ada Lovelace Institute / Alan Turing Institute, November 2024).

From 5 February 2026 the amended UK GDPR uses Articles 22A–22D for significant solely automated decisions. An appropriate lawful basis other than recognised legitimate interest is required; special category decisions remain restricted and safeguards include information, representations, human intervention and a way to contest the decision. The applicable law cannot be inferred from public acceptance percentages.

What safeguards do people want for AI and their data?

In the March 2026 survey, 51% selected human review or checking among the safeguards that would increase trust. Forty-seven per cent selected information about challenging, correcting or appealing a decision among their top three information priorities. Twenty-four per cent said nothing would increase their trust. These were multiple-response questions, not a single ranked choice.

Transparency demands are near-universal: 93% say it is important that AI-generated content is clearly labelled, including 80% who say very important (March 2026). Combined with the 53% who want mandatory chatbot disclosure, the public's message to organisations is consistent — tell people when they are dealing with AI, and give them a human route out.

The Ada Lovelace and Alan Turing institutes’ 2024/25 research found 72% saying laws and regulation would make them more comfortable with AI, compared with 62% in its earlier survey. That measures reported attitudes to safeguards; it does not prove that a particular law or policy increases trust.

For organisations, the practical reading is straightforward: human review, clear labelling, a visible appeal route and staff who understand the data protection rules that govern all three. Where AI use goes wrong and people complain to the regulator, those complaints show up in the ICO's caseload — tracked on our ICO complaints statistics page.

Frequently asked questions

How many UK businesses use AI?

ONS estimated 30% of UK businesses used at least one AI technology in September 2026, including 51% of businesses with 250 or more employees. UKBDS 2026, using a different question and the narrower population handling digitised data, found 41%, rising to 82% of large businesses.

What proportion of UK adults have used an AI chatbot?

In March 2026, 74% of UK adults said they had ever interacted with an AI chatbot in a question including customer-service bots and virtual assistants. DSIT’s July–August 2024 online survey found 60% had used chatbots in the previous three months. These questions and timeframes are not directly comparable.

Are UK adults comfortable with AI making decisions about them?

The March 2026 ICO survey found 76% concerned about computers making decisions about people. AI acceptability varied by use: 52% accepted fraud detection, 15% hiring and 14% benefits decisions. Separate questions explicitly excluding human involvement had different results.

Do UK businesses have AI usage policies?

UKBDS 2026 chapter 3 reports 17% of AI-using businesses had a policy or guidance: 5% formal written and 12% informal. Formal policies were reported by 56% of large AI-using businesses and 3% of AI-using sole traders. A policy’s presence is not a compliance rating.

Does regulation make people more comfortable with AI?

In the Ada Lovelace and Alan Turing institutes’ 2024/25 study, 72% said laws and regulation would make them more comfortable with AI. In the separate March 2026 ICO survey, 51% chose human review among safeguards that would increase trust. These are stated attitudes, not measured effects of regulation.

Sources & references

Most teams now use AI tools that touch personal data — but few have the training or policies to do it within UK GDPR. Close that gap.

Explore the GDPR & Data Protection Course →
Mark McShane
Mark McShane
Health & Safety Training Specialist, Online CPD Academy

Mark writes about data protection, workplace compliance and accredited online training for GDPR & Data Protection Course, part of Online CPD Academy.