AI Search Adoption Statistics 2026: Who Uses AI, How Often, and Whether They Stick

Three years after ChatGPT arrived, the argument has moved on. Nobody is asking whether people will use AI tools to find things out. The question now is who, how often, and whether they stick. A headline saying "half of adults use AI" hides an enormous spread: a 24 year old who opens ChatGPT six times a day and a 68 year old who tried it once at Christmas are both counted in the same number. What follows is the adoption data published in 2025 and 2026 by statistical agencies, regulators and pollsters, with a UK focus section. The subject is people and businesses, and the gap between having tried something and actually depending on it.

Key AI search adoption statistics for 2026

• 49% of US adults now use AI chatbots, up from 33% in 2024, according to the Pew Research Center [1]
• 24% of US adults use a chatbot daily, meaning roughly half of all users are daily users, per Pew Research Center [1]
• 35% of UK adults now use generative AI tools weekly, and 58% have used them at all, on Ipsos fieldwork for the BBC [2]
• 35% of UK businesses with 10 or more employees reported using at least one AI technology by June 2026, up from around 12% in September 2023, per the Office for National Statistics [3]
• 1.6 AI technologies are used by the average adopting UK business, barely up from 1.4 in 2023, per the ONS [3]
• 54.3% of UK firms surveyed in January 2026 were using AI in some form, up from 23% in 2023, according to British Chambers of Commerce research with the University of Essex [4]
• 52% of US employees used AI at work in Q2 2026, with 15% using it daily, per Gallup [5]
• 32% of UK online adults visited a major AI chatbot in June 2025, around 15.8 million people, on Ipsos iris data cited by Ofcom [6]
• 66% of adults across 21 countries had used an AI tool in the previous twelve months, per the Google/Ipsos multi-country survey [7]
• 10% of people globally now use AI chatbots for news weekly, up from 7%, but only 4% in the UK, per the Reuters Institute [8]
• 655 million unique visitors reached AI chatbot sites monthly in the year to May 2026, up 57% year on year, per Similarweb [9]
• 67% of US non-users say they are unlikely to start using chatbots in the next twelve months, per Pew Research Center [10]

How many people actually use AI chatbots in 2026?

Just under half of US adults, and the number is still climbing. Pew Research Center's February 2026 survey of 5,119 US adults found 49% use AI chatbots, up from 33% in 2024 [1]. Globally the figure is higher on some measures.

The Google/Ipsos survey of roughly 21,000 adults across 21 countries, fielded September to October 2025, put usage at 66% in the previous twelve months, up 18% on 2024 and 28% on 2023 [7].

Those two numbers are not in conflict, they measure different things. Pew asks about current use of chatbots specifically. Google/Ipsos asks about any AI tool in the last year, a much wider net that catches anyone who once used a translation feature or an AI photo filter.

The looser the definition and the longer the recall window, the bigger the number, and vendors quote the biggest number they can find.

Similarweb's panel data counts behaviour rather than claims: 655 million unique visitors to AI chatbot properties monthly in the year to May 2026, growing 57% year on year, against visit growth of 70% [9]. Visits growing faster than visitors is the interesting bit. Existing users are going back more often, a shift I dig into more in the AI search behaviour statistics.

What proportion use AI every day rather than occasionally?

Roughly a quarter of US adults, which is about half of all users. Pew found 24% of US adults use chatbots daily, with another 25% using them several times a week or less [1]. That is an unusually high daily share for a technology this young.

Compare that with the workplace. Gallup's Q2 2026 survey of 22,573 employed US adults found 52% had used AI at work, but only 15% used it daily and 30% a few times a week or more [5].

Source: Gallup [5]

Read that table from right to left. Total use has roughly doubled in three years, but daily use has nearly quadrupled, so growth is coming as much from existing users deepening their habit as from new arrivals.

The daily column is the one that changes behaviour. Occasional users still default to Google. Daily users build a reflex, and reflexes are what move traffic.

Who is adopting AI, by age?

Adoption falls sharply with age, but the fastest growth is now in older groups. Pew's age data shows every bracket rising between 2024 and 2026, with the middle-aged catching up quickest [11].

Source: Pew Research Center [1] [11]

The row to watch is 30-49. They are not the most likely to have tried AI, but they are the most likely to use it daily, at 34%. That is the group with jobs, mortgages, children and no spare time, exactly the profile of someone who adopts a tool because it saves effort rather than because it is fun.

Ipsos found something similar in the UK, reporting the sharpest growth among the over 55s, a 250% increase in that group [2]. Early adoption stories are always about the young. Mass adoption stories are about everyone else arriving late and then not leaving.

How does UK adoption compare with other countries?

The UK is a solid mid-table adopter of AI tools generally, but a notable laggard when it comes to trusting AI for information. Ipsos fieldwork for the BBC, conducted in spring 2025 among 2,000 UK adults aged 16 to 75 and published in February 2026, found 58% had used generative AI tools and 35% used them weekly [2].

Ofcom's own numbers are more conservative. Its November 2025 discussion paper cited 2024 survey data showing 41% of UK internet users aged 16 and over had used a generative AI tool in the past year, with only 18% having used one in the past month, rising to 78% past-year use among 16 to 24 year olds [6].

On behavioural data, Ipsos iris measurement cited by Ofcom put combined monthly reach of the main AI chatbots at 15.8 million UK online adults in June 2025, or 32%, with ChatGPT alone reaching 13.0 million (26%) [6].

The outlier is news. The Reuters Institute's Digital News Report 2026, covering 48 markets with around 2,000 respondents each, found weekly AI chatbot use for news at 10% globally, up from 7%, and 16% among under 35s. The UK sat at 4%, the lowest of any market surveyed, with trust in AI answers at 6% against a global 20% [8] [12].

British scepticism is measurable. That does not mean the tools are unused here, it means UK users are more likely to treat AI as a utility for tasks than as an authority on facts. If you are optimising for AI visibility in the UK, that scepticism is your operating environment.

How many businesses have adopted AI, and does size matter?

Size matters enormously, and the UK official data is unusually good on this. The ONS reports that use of at least one AI technology among businesses with 10 or more employees rose from around 12% in September 2023 to around 35% by June 2026 [3].

Beneath that, the spread by headcount is wide: 28% of businesses with 0 to 9 employees use AI, against 49% of those with 250 or more [3].

Source: ONS, June 2026 [3]

The British Chambers of Commerce study with the University of Essex, based on 672 responses to its January 2026 Business Outlook Survey, found a higher headline of 54.3% currently using AI with a further 17.5% intending to, up from 23% in 2023 [4]. Its sector spread ran from 77.8% in IT and digital and 70.0% in creative and media down to 44.5% in manufacturing and 30.5% in other services [4].

The BCC number sits well above the ONS number for the same period. That is a sampling story, not a contradiction. Chamber members are self-selecting and more engaged, whereas the ONS draws from a statistical survey covering the whole business population.

When two credible sources disagree by twenty points, the honest answer is that UK business AI adoption is somewhere between one in three and one in two firms, and the direction is up.

Globally, McKinsey's survey of 1,491 participants across 101 nations, published March 2025 with fieldwork in July 2024, put organisational adoption at 78% using AI in at least one business function and 71% regularly using generative AI [13]. Executive panels almost always report far higher adoption than population-wide business surveys, because the population of businesses is mostly very small businesses.

Is business adoption getting deeper, or just wider?

Wider. Considerably wider, and barely deeper. This is the most underrated finding in the 2026 UK data.

The ONS reports that the average number of AI technologies used per adopting business rose only from around 1.4 to 1.6 between 2023 and 2026 [3]. Nearly three times as many firms are using AI, and they are using it in almost exactly the same shallow way.

The technology mix explains why. Large language models lead at 18% of businesses, followed by visual content creation at 16%, machine learning for data processing at 12%, image processing at 6% and robotics at 2% [3].

Most UK business AI adoption, in other words, is a person with a chatbot tab open. That is a real productivity gain and it is also nothing like the operational transformation the vendor decks describe.

Gallup's April 2026 survey of 23,717 employed US adults found frequent AI use at 67% among leaders but 46% among individual contributors [14]. Adoption is concentrated at the top of organisations, where the enthusiasm is, rather than in the roles where volume work actually happens.

Why do people who try AI stop using it?

Because most non-users are not hesitant, they are uninterested. Pew's data is blunt on this: 51% of US adults never use chatbots, and among those non-users, 83% cite lack of interest as a reason, 79% privacy concerns and 76% accuracy concerns [10].

More significantly for anyone forecasting adoption, 67% of non-users say they are unlikely to start within the next twelve months, including 40% who say "not at all" likely. Only 5% expect to definitely start soon [10].

Non-use is heavily age-weighted: 77% of over 65s do not use chatbots, against 34% of 18 to 29 year olds [10].

Gallup found the workplace version of the same thing. Among employees who do not use AI, 46% simply prefer their current way of working, 43% cite data privacy, 43% object on ethical grounds and 39% do not believe AI can help with their job [14].

The remaining non-adopters are a harder sell than the ones who came before them. Growth from here will be slower per percentage point, and plans that assume a straight line to 90% adoption are extrapolating from the easy half of the curve.

Are people using AI to search, or to do something else?

Search is the single biggest use case, but it is not the majority use case. Pew found 42% of US adults use chatbots for searching for information, ahead of work tasks at 38% among employed adults, entertainment at 25%, image or video creation at 24% and medical advice at 20% [1].

Ofcom's UK data broadly matched, with 36% citing finding information or content as a reason for using generative AI, ahead of exploring the technology at 30%, summarising text at 24% and seeking advice at 23% [6].

Passive exposure to AI in search is far larger than active use of chatbots. Ofcom reports that about 30% of searches now show AI overviews and more than half of adults (53%) say they see these summaries often, while Google Search still reaches 82% of UK adults and handles around 3 billion UK searches a month [15].

That is the honest shape of AI search adoption in 2026. A minority actively choose an AI tool to look something up. A majority get AI answers whether they asked for them or not, a dynamic the wider AI search statistics set out in full.

A commerce layer is forming on top. Adobe's January 2026 analysis reported that 47% of surveyed consumers trust AI recommendations and 81% said AI improved their shopping experience [16].

How to read these numbers

Survey adoption and measured adoption are different things. Pew, Ipsos, Gallup and the BCC all ask people what they do, which invites over-claiming on socially desirable behaviour and under-claiming when respondents do not recognise the label "AI". Similarweb and Ipsos iris measure devices and sessions, which catches accidental visits and misses in-app usage.

Recall windows change everything. "Ever used", "used in the last twelve months", "used in the last month" and "use weekly" can differ by forty percentage points for the same population.

Business surveys are dominated by whoever answers them. The ONS covers the full business population including microbusinesses, which drags the average down. Executive panels cover large firms with AI budgets, which pushes it up. Neither is wrong.

Fieldwork dates lag publication dates, sometimes badly. The Ipsos UK figure of 58% was published in February 2026 but collected in spring 2025, which in this market is close to a generation ago.

Finally, several figures here are single-market. US chatbot behaviour is not UK chatbot behaviour, as the Reuters Institute's 4% UK news figure demonstrates [8].

What this means going into 2027

Expect adoption breadth to slow and depth to accelerate. The people left outside are actively uninterested, and 67% of US non-users say they will not join within a year [10]. The growth story from here is frequency, not headcount.

Watch the 30-49 bracket rather than Gen Z. They already show the highest daily use at 34%, they hold the buying decisions, and they are the group whose habits will determine whether AI-mediated discovery becomes commercially material [11].

UK businesses will keep adopting shallowly unless something changes. At 1.6 technologies per adopting firm, most are one chatbot subscription away from where they started, and the ONS trend gives little sign of that deepening on its own [3].

For UK-facing marketers, plan for exposure rather than preference. Most British users are not choosing AI over search, they are receiving AI answers inside search, and UK trust in AI-supplied information remains the lowest measured anywhere [8] [15].

And treat any 2027 forecast quoting a single adoption percentage with suspicion. The spread between "has tried" and "uses weekly" is now the whole ballgame, and anyone collapsing it into one number is selling something.

If you want to work out what these adoption curves mean for your own brand's visibility across the AI engines, and a concrete plan to act on it, you can book a call with me and we will go through your numbers together.