AI Search Statistics: What the Named Sources Actually Say
Table of Contents
AI Search Statistics at a Glance
Last verified: 17 September 2026. Ofcom reports that 54 percent of UK adults now use AI tools such as ChatGPT, Copilot or Gemini, up from 31 percent in 2024. UKOM Ipsos iris measured 29.6 million people using AI tools in the UK in July 2026, an increase of 3.1 million in three months. Similarweb data reported by TechCrunch puts AI Overviews on 43 percent of United States Google searches as of May 2026, against roughly 15 percent a year earlier.
Pew Research Center found that users clicked a traditional search result on 8 percent of visits where an AI summary appeared, compared with 15 percent where it did not. Every one of the AI search statistics on this page carries a named source, a publication date and the date it was last checked.
This page collects AI search statistics that can be traced back to a named organisation and a dated publication. It exists because most AI search usage statistics circulating online are copied between blogs until the original measurement, its sample and its date are lost. The figures are grouped by what they actually measure: adoption, assistant market activity, AI Overviews statistics on coverage, click-through effects and business impact. If you are working out what these numbers mean for your own visibility rather than just quoting them, our guide to AI SEO covers the practical side of the same shift.
A figure that cannot be sourced does not appear in these AI search statistics. Where two credible sources disagree, both are shown and the disagreement is explained rather than hidden behind an average.
What Counts as AI Search?
AI search covers two different behaviours that are often merged into one number, which is the single biggest reason published AI search statistics conflict with each other.
The first is generated answers inside a traditional search engine. Google’s AI Overviews and AI Mode sit here. The user is still searching Google, but the result page opens with a written answer rather than a list of links.
The second is assistant use, where someone asks ChatGPT, Gemini, Copilot, Claude or Perplexity a question directly and never opens a search engine. These sessions leave no search engine query data at all.
Statistics describing one behaviour are routinely presented as describing the other. When you see a claim that some percentage of searches “now happen in AI”, check which of the two it measured. The AI search usage statistics below keep the two apart throughout.
AI Search Statistics: Adoption in the UK
The UK AI search statistics in this section come from two named panels that publish their fieldwork dates, which makes them the most quotable figures on the page.
54 percent of UK adults use AI tools such as ChatGPT, Copilot or Gemini.
Up from 31 percent in 2024. Based on Ofcom’s Adults’ Media Use and Attitudes tracker, with fieldwork running from late September to late November 2025.
Source: Ofcom, Adults’ Media Use and Attitudes 2026, published 1 April 2026. Checked 17 September 2026.
79 percent of UK adults aged 16 to 24 use AI tools, and 74 percent of those aged 25 to 34.
Adoption falls with age, which is why a single national figure understates exposure among younger audiences.
Source: Ofcom’s 2026 research into UK adults’ media habits, published 1 April 2026. Checked 17 September 2026.
75 percent of online UK adults read AI-generated search summaries at least sometimes.
Among UK adults who do not use AI chatbots at all, 54 percent still read AI search summaries at least sometimes. Exposure does not require deliberate use of an assistant.
Source: Ofcom’s Adults’ Media Use and Attitudes 2026 report, published 2 April 2026. Checked 17 September 2026.
29.6 million people in the UK used AI tools in July 2026.
An increase of 3.1 million users in three months. Measured across PC, laptop, smartphone and tablet for all aged 15 and over. The measured category excludes AI built into other products such as Google Search and WhatsApp.
Source: UKOM Ipsos iris data published by IAB UK, released July 2026. Checked 17 September 2026.
More than a third of UK people aged 55 and over used AI tools in July 2026.
Up from just over a fifth in July 2025. Growth in UK AI adoption is now driven by older age groups rather than the youngest.
Source: IAB UK’s AI usage figures from UKOM Ipsos iris, released July 2026. Checked 17 September 2026.
57 percent of UK adults aware of AI would trust a news story written by AI less than one written by a person.
A further 27 percent said they would trust it about the same, and 7 percent said they would trust it more. Adoption and trust are moving at different speeds.
Source: Ofcom’s 2026 media use report, published 2 April 2026. Checked 17 September 2026.
UK AI Tool Audience Over Time
The clearest UK series available uses one consistent methodology, so these AI search usage statistics are comparable with each other in a way that figures pulled from different vendors are not.
| Month Measured | UK People Using AI Tools |
|---|---|
| July 2025 | 20.2 million |
| September 2025 | 21.1 million |
| October 2025 | 22.5 million |
| January 2026 | 24.0 million |
| April 2026 | 26.4 million |
| July 2026 | 29.6 million |
Source for all rows: UKOM Ipsos iris Online Audience Measurement Service, published by IAB UK. All figures cover people aged 15 and over using PC, laptop, smartphone or tablet. Checked 17 September 2026.
AI Search Usage Statistics: Assistant Market Activity
ChatGPT reached 900 million weekly active users in February 2026.
Up from 800 million in October 2025 and 400 million in February 2025. This remains the most recent weekly figure OpenAI has published itself, which matters when you see larger numbers quoted without attribution.
Source: OpenAI, reported by TechCrunch on 27 February 2026. Checked 17 September 2026.
OpenAI reported 50 million consumer subscribers and more than 9 million paying business users in February 2026.
Disclosed alongside the weekly active user figure. Paid adoption remains a small fraction of total usage.
Source: OpenAI, via TechCrunch’s February 2026 report. Checked 17 September 2026.
The Gemini app passed 1 billion monthly active users in August 2026.
Google reported 400 million monthly users in May 2025 and more than 950 million at its second quarter 2026 earnings. Gemini is Google’s fourteenth product to reach a billion users.
Source: Google, announced 11 August 2026. Checked 17 September 2026.
63 percent of Gemini app users now interact with it by voice.
Google also reported that one in five Gemini Live interactions goes beyond voice, using live camera or screen sharing. Assistant use is drifting away from typed text, which affects how queries are phrased.
Source: Google’s August 2026 Gemini milestone announcement. Checked 17 September 2026.
More than 100 million Gemini app users are on iOS.
Reported at the same billion-user milestone. Usage is not confined to Android devices where Gemini ships by default.
Source: Google, reported by TechCrunch on 11 August 2026. Checked 17 September 2026.
ChatGPT Usage Statistics: UK Audience and Share of Attention
The ChatGPT usage statistics UK marketers most often need are audience reach and share of time, and both come from the same panel rather than from OpenAI.
76 percent of all time UK users spent on AI tools in January 2026 went to ChatGPT.
Down from 84 percent in October 2025. Share of attention is falling even as total AI use rises, which is the more useful of the available ChatGPT usage statistics for the UK because it measures engagement rather than reach.
Source: UKOM Ipsos iris figures via IAB UK, released January 2026. Checked 17 September 2026.
ChatGPT reached a UK monthly audience of 16 million in September 2025.
Up from 15.1 million in July 2025, when Microsoft Copilot was the second most used AI tool in the UK at 2.5 million people. This remains the most recent UK audience figure published for ChatGPT specifically by that measurement service.
Source: IAB UK, citing UKOM Ipsos iris, released November 2025. Checked 17 September 2026.
Claude recorded the largest monthly UK audience growth of any AI tool between January and April 2026, at 232 percent.
Growth rate and absolute audience are different measures. A tool can lead on growth from a small base while share of time stays concentrated elsewhere.
Source: UKOM Ipsos iris data released by IAB UK, April 2026. Checked 17 September 2026.
AI Overviews Statistics: How Often Generated Answers Appear
AI Overviews appeared on 43 percent of United States Google searches in May 2026.
Up from roughly 15 percent a year earlier. Measured by Similarweb panel data. The published materials do not state the device mix or vertical weighting, so treat the trend as solid and the exact figure as an estimate.
Source: Similarweb, reported by TechCrunch on 27 July 2026. Checked 17 September 2026.
Google AI Mode visits grew from 126 million in June 2025 to 279 million in May 2026.
AI Mode visits and AI Overviews coverage measure different things. The first counts people choosing a conversational surface, the second counts how often a generated answer is served.
Source: Similarweb data reported by TechCrunch’s July 2026 analysis of AI search. Checked 17 September 2026.
58 percent of tracked users ran at least one Google search in a single month that returned an AI summary.
Measured across 900 United States adults during March 2025. Reach across users is higher than coverage across queries, and the two are frequently confused in published AI Overviews statistics.
Source: Pew Research Center, published 22 July 2025. Checked 17 September 2026.
Query length predicts whether an AI summary appears: 8 percent of one or two word searches produced one, against 53 percent of searches of ten words or more.
Question-based searches beginning with who, what or why produced an AI summary 60 percent of the time. Conversational phrasing is the strongest trigger available in the public data.
Source: Pew Research Center’s analysis of Google AI summaries, published 22 July 2025. Checked 17 September 2026.
Coverage estimates from different measurement firms sit several percentage points apart because each tracks a different sample of queries. Any AI Overviews statistics quoted as a single precise national percentage are reporting one vendor’s panel rather than a Google disclosure. Google has not published a coverage figure of its own.
AI Overviews Statistics: Click-Through Effects
The click-through AI search statistics below come from tracked browser behaviour rather than from modelled estimates, which is why they carry more weight than most published alternatives.
| Behaviour Measured | Search Page With an AI Summary | Search Page Without One |
|---|---|---|
| Clicked a traditional search result | 8% of visits | 15% of visits |
| Clicked a link inside the AI summary | 1% of visits | Not applicable |
| Ended the browsing session | 26% of visits | 16% of visits |
| Continued searching Google for brand content | 32% of visits | 35% of visits |
Source for all rows: Pew Research Center, based on 900 United States adults and nearly 69,000 Google searches during March 2025, published 22 July 2025. Checked 17 September 2026.
Only 1 percent of visits to a search page carrying an AI summary produced a click on a source cited inside that summary.
Citation inside an AI answer and traffic from that answer are separate outcomes. A page can be cited constantly and receive almost no visits from it.
Source: Pew Research Center’s browsing study of AI summaries, published 22 July 2025. Checked 17 September 2026.
Google disputes the methodology behind the published click-through declines.
The company has stated that organic click volume from search to websites has remained consistent year on year, and has criticised sample sizes and isolated analysis periods. Google has not released comparable click-through data of its own, so the dispute cannot currently be settled from public sources.
Source: Google’s public response, reported alongside the findings by Pew Research Center and subsequent trade coverage, July 2025 onward. Checked 17 September 2026.
The practical reading of these AI Overviews statistics is that visibility inside a generated answer has to be measured separately from clicks, because the two have separated. That separation is why measurement approaches such as an AI visibility audit now sit alongside conventional rank tracking rather than replacing it.
AI Search Statistics: Business Impact
Adoption and coverage figures describe exposure. The AI search statistics in this section describe what happens commercially once someone arrives from a generated answer.
Traffic from AI sources to United States retail sites grew 393 percent year on year in the first quarter of 2026.
March 2026 alone was up 269 percent year on year. The preceding holiday period of November and December 2025 ran 693 percent ahead of the year before. Based on more than one trillion visits to United States retail sites.
Source: Adobe Analytics, published April 2026. Checked 17 September 2026.
AI-sourced traffic converted 42 percent better than non-AI traffic in March 2026.
Twelve months earlier, in March 2025, the same channel converted 38 percent worse. The reversal happened inside a single year, which is a strong argument for rechecking any conversion benchmark older than two quarters.
Source: Adobe’s first quarter 2026 AI traffic analysis. Checked 17 September 2026.
AI-sourced retail traffic grew 138 percent year on year in May 2026, and 1,324 percent since October 2024.
October 2024 is when Adobe began tracking these referrals, so it functions as the baseline rather than a natural starting point.
Source: Adobe Analytics, published 17 June 2026. Checked 17 September 2026.
AI traffic converted 54 percent better than non-AI sources in May 2026.
Adobe also reported that 80 percent of consumers who use AI for shopping say they use it more than they used to, and 79 percent feel more confident in purchases made with AI assistance.
Source: Adobe’s June 2026 report on AI-sourced traffic. Checked 17 September 2026.
39 percent of surveyed consumers have used AI for online shopping, and 85 percent of those said it improved the experience.
Engagement rate from AI sources ran 12 percent higher than non-AI traffic in March 2026 once a visitor landed on a retail site.
Source: Adobe Analytics survey data, published April 2026. Checked 17 September 2026.
Read together, the click-through and business impact numbers describe one pattern rather than two contradictory ones. Fewer visits arrive, and the visits that do arrive are further along in a decision. That is the mechanism behind generative engine optimisation as a distinct discipline: the goal shifts from maximising sessions to being the source a system reaches for.
How These AI Search Statistics Were Sourced and Verified
Every figure above was checked against its original publisher on 17 September 2026, not against a secondary article quoting it. Where the original was a press release or a report PDF, that document is what the link points to.
Sources were chosen against four tests. The publisher had to be identifiable. The measurement date had to be stated. The sample or panel had to be described well enough to know what the number covers. And the organisation had to have a reason to measure accurately beyond marketing its own services.
That test excludes a large share of the AI search statistics in circulation. It rules out figures with no named originator, figures whose only citation is another blog, and figures presented without the month they were measured. It also rules out several widely quoted numbers that trace back to vendor studies with no published methodology.
Three limitations are worth stating plainly rather than burying. Coverage percentages come from commercial panels, not from Google, so they carry an unknown margin. Click-through findings are contested by Google and have not been resolved with public data from either side. And UK-specific measurement of AI Overviews coverage does not yet exist at national level, so the AI Overviews statistics here are United States measurements and are labelled as such.
This page is reviewed every quarter. Each review rechecks whether a newer measurement has replaced the one shown, updates the verified date, and removes any figure whose source has been withdrawn or superseded. The next scheduled review of these AI search statistics is December 2026. Where a number is replaced, the previous figure and its date are kept in the surrounding text so the direction of travel stays visible.
Frequently Asked Questions About AI Search Statistics
What counts as AI search in these statistics?
Two behaviours are counted, and they are kept apart throughout this page. The first is a generated answer inside a conventional search engine, such as Google’s AI Overviews or AI Mode. The second is a direct question to an assistant such as ChatGPT or Gemini, where no search engine is involved at all.
How are AI search usage statistics actually measured?
Three methods produce almost all published AI search usage statistics. Panel measurement recruits a sample of people and records what they do, which is how Ofcom, UKOM Ipsos iris and Pew Research Center work. Query tracking observes result pages at scale to see which features appear, and platform disclosure is the companies reporting their own user counts, which is accurate on reach but tells you nothing about behaviour.
Why do AI search statistics differ so much between studies?
Usually because they are measuring different denominators and calling them the same thing. Coverage of queries, reach across users and share of sessions all get reported as a single percentage of “searches”. Sample composition, country, device mix and measurement month then move the number further apart.
I keep seeing different AI Overviews statistics for how many searches show a generated answer. Which one is right?
None of them is wrong exactly, and that’s the frustrating part. Commercial measurement firms track their own keyword panels, so a five point gap between two vendors reflects two different samples rather than one of them being mistaken. Google has not published its own coverage figure, so there is no authoritative number to check them against, and anyone presenting one precise percentage as settled fact is overstating what the data supports.
Just give me one AI search statistic I can put in a client report without it being wrong next month.
The Ofcom adoption figure is the safest, because it comes from a UK regulator, states its fieldwork dates and is updated annually rather than monthly. Quote it as “54 percent of UK adults use AI tools, according to Ofcom’s Adults’ Media Use and Attitudes 2026 report” and it will hold up under scrutiny. Avoid the fast-moving coverage and click-through percentages in any document that needs to stay accurate for more than a quarter.
Where do the ChatGPT usage statistics UK marketers quote actually come from?
Almost none of them come from OpenAI, which reports global user numbers and does not break them out by country. The ChatGPT usage statistics UK teams can legitimately cite come from UKOM Ipsos iris panel measurement, published through IAB UK, covering monthly audience and share of time spent. If you see a UK-specific ChatGPT user count with no panel named behind it, it has almost certainly been estimated from global figures rather than measured.
Are there UK-specific AI search statistics, or is everything American?
UK adoption data is good, and comes from Ofcom and from UKOM Ipsos iris via IAB UK. UK-specific ChatGPT usage statistics exist at the level of audience and time share. What does not yet exist at national level is UK measurement of AI Overviews coverage or click-through effects, so those figures on this page are United States measurements and are labelled that way.
Do AI Overviews definitely reduce traffic to websites?
The independent measurement points consistently in that direction, and Google disputes it. Pew Research Center tracked real browsing and found clicks roughly halved when an AI summary appeared, while Google has said organic click volume to websites has held steady year on year and has criticised the methodology. Google has not released data that would settle the question, so the honest answer is that the evidence available to the public favours a decline without proving it.
How often is this page of AI search statistics updated?
Quarterly. Each review rechecks every figure against its original source, updates the verified date at the top, replaces superseded numbers and removes anything whose source has been withdrawn. The next scheduled review is December 2026.
Can I cite these AI search statistics?
Yes, and citing the original publisher is better than citing this page. Each block names the organisation, the publication date and the date the figure was last checked, so you can go straight to the source and confirm it is still current. If you find a figure in these AI search statistics that no longer matches its source, the page is wrong and we would rather fix it than have it quoted.