Half of American adults now use AI chatbots, and 60% encounter AI-written summaries in search whether they sought them out or not. Adoption reached 49%, up 16 points in two years, with ChatGPT alone at 44%. But usage is running well ahead of confidence: 63% say AI is advancing too fast, 71% expect it to make their personal information less secure, and 67% have little or no faith in government to regulate it. The figures below come from Pew Research Center’s survey of 5,119 US adults.
TL;DR
- 49% of US adults use AI chatbots, up from 33% in 2024 (Pew)
- 24% use chatbots daily (Pew)
- 12% use them several times a day and 4% almost constantly (Pew)
- 51% do not use chatbots at all (Pew)
- ChatGPT reaches 44%, up from 18% in 2023 (Pew)
- Gemini reaches 24% and Copilot 17% (Pew)
- Searching for information is the top use at 42% (Pew)
- 60% read AI-generated search summaries (Pew)
- 63% say AI is advancing too fast (Pew)
- 71% expect AI to make their personal information less secure (Pew)
- 67% have little or no confidence in government regulation (Pew)
- 40% expect AI to have a negative societal impact within 20 years (Pew)
- Smartwatches reach 37% and smart speakers 35% (Pew)
1. Adoption Crossed the Halfway Mark
The headline figure sits almost exactly at the midpoint, which makes the trend more informative than the level. 49% of US adults use AI chatbots, up from 33% two years earlier, a 16-point rise that works out to roughly 8 points a year.
Depth matters as much as reach here. Of the 49%, roughly half are daily users: 24% use a chatbot every day, including 12% several times a day and 4% almost constantly. The remaining 25% use them several times a week or less. That distribution describes a technology that has found genuinely heavy users rather than one people tried once. At the same time 51% still do not use chatbots at all, so the common framing of AI as universally adopted describes the trajectory rather than the present.
| Metric | Value | Source |
|---|---|---|
| Use AI chatbots | 49% | Pew Research Center |
| Two years earlier | 33% | Pew Research Center |
| Change | up 16 points | Pew Research Center |
| Implied annual pace | approx. 8 points | Derived from Pew figures |
| Daily users | 24% | Pew Research Center |
| Several times a day | 12% | Pew Research Center |
| Almost constantly | 4% | Pew Research Center |
| Several times a week or less | 25% | Pew Research Center |
| Non-users | 51% | Pew Research Center |
Source: Pew Research Center, Americans and AI 2026.
2. ChatGPT Dominates but Is Not Alone
The brand-level data shows one clear leader and a competitive field behind it. ChatGPT reaches 44% of US adults, up from 34% a year earlier and 18% in 2023, with Gemini at 24%, Copilot at 17%, Meta AI at 14%, Grok at 8%, Claude at 6%, and Character.ai at 3%.
Those shares sum to 116%, well above the 49% who use any chatbot, which means multi-tool use is the norm rather than the exception. That has a practical implication often missed: these are not mutually exclusive market shares and cannot be read as a competitive split. ChatGPT at 44% against overall chatbot use at 49% is the more revealing comparison, since it means roughly nine in ten chatbot users touch ChatGPT. Whatever competition exists is happening around a product that is close to synonymous with the category for most people.
| Metric | Value | Source |
|---|---|---|
| ChatGPT | 44% | Pew Research Center |
| ChatGPT a year earlier | 34% | Pew Research Center |
| ChatGPT in 2023 | 18% | Pew Research Center |
| Gemini | 24% | Pew Research Center |
| Copilot | 17% | Pew Research Center |
| Meta AI | 14% | Pew Research Center |
| Grok | 8% | Pew Research Center |
| Claude | 6% | Pew Research Center |
| Character.ai | 3% | Pew Research Center |
| Sum of named tools | 116% | Derived from Pew figures |
| Share of chatbot users touching ChatGPT | approx. 90% | Derived from Pew figures |
Accuracy context sits in our AI hallucination statistics. Source: Pew Research Center, what Americans think AI is.
3. What People Actually Do With It
The use-case breakdown is more mundane than either the enthusiasm or the alarm suggests. Searching for information leads at 42%, work tasks reach 38% among employed adults, entertainment 25%, and image or video creation 24%.
The lower half of the list is where the interesting signals are. Medical advice and diet or fitness each reach 20%, which is a substantial share of adults taking health guidance from a system with no clinical accountability. News sits at 13%, emotional support at 10%, and companionship at 4%. The companionship figure is worth noting precisely because it is small: the cultural conversation about AI relationships describes a real but narrow behaviour, roughly one adult in 25, while the dominant use is a slightly better search box.
| Metric | Value | Source |
|---|---|---|
| Searching for information | 42% | Pew Research Center |
| Work tasks, employed adults | 38% | Pew Research Center |
| Entertainment | 25% | Pew Research Center |
| Image or video creation | 24% | Pew Research Center |
| Medical advice | 20% | Pew Research Center |
| Diet or fitness | 20% | Pew Research Center |
| News | 13% | Pew Research Center |
| Emotional support | 10% | Pew Research Center |
| Companionship | 4% | Pew Research Center |
Workplace context sits in our AI impact on jobs statistics. Source: Pew Research Center, Americans and AI 2026.
4. Passive Exposure Beats Deliberate Use
One figure in this dataset exceeds every measure of intentional chatbot use. 60% of adults say they read AI-generated summaries in search results, against 49% who use chatbots at all.
That gap is the most consequential finding here. The largest channel through which Americans encounter generative AI output is one they did not choose, did not open, and in many cases may not identify as AI. It arrives above the search results they were already looking for. Every debate about adoption rates, tool preference, and trust in chatbots concerns a smaller population than the one being served AI-written text daily by default. Measures of AI adoption built on deliberate use systematically understate exposure, and the trust findings in the next section apply to a product most people are consuming without opting in.
| Metric | Value | Source |
|---|---|---|
| Read AI-generated search summaries | 60% | Pew Research Center |
| Use AI chatbots at all | 49% | Pew Research Center |
| Gap | 11 points | Derived from Pew figures |
| Use ChatGPT | 44% | Pew Research Center |
| Gap versus ChatGPT | 16 points | Derived from Pew figures |
| Nature of the exposure | passive, default-on | Derived |
| Implication for adoption metrics | deliberate-use figures understate exposure | Derived |
| Daily chatbot users for comparison | 24% | Pew Research Center |
Information-quality context sits in our misinformation statistics. Source: Pew Research Center, why don’t people use chatbots.
5. Adoption Without Confidence
The sentiment data runs firmly against the usage data. 63% say AI is advancing too fast, 71% expect it to make their personal information less secure, and 67% have little or no confidence in the US government to regulate it effectively, up from 62% two years earlier.
Confidence in the private sector is barely better, with 59% expressing little or no confidence that companies will act responsibly. On impact, 40% expect AI to have a negative effect on society over the next 20 years while 31% expect a negative effect on themselves personally, a nine-point gap that shows a familiar pattern of people judging collective risk more harshly than individual risk. Younger adults are the most pessimistic about society, at 48% among those aged 18 to 29 against 39% for 30 to 49 and 37% for 50 and over, which cuts against the assumption that unease about AI tracks unfamiliarity.
| Metric | Value | Source |
|---|---|---|
| AI advancing too fast | 63% | Pew Research Center |
| Expect personal information less secure | 71% | Pew Research Center |
| Little or no confidence in government regulation | 67% | Pew Research Center |
| Same figure two years earlier | 62% | Pew Research Center |
| Little or no confidence in corporate responsibility | 59% | Pew Research Center |
| Expect negative societal impact in 20 years | 40% | Pew Research Center |
| Expect negative personal impact | 31% | Pew Research Center |
| Societal versus personal gap | 9 points | Derived from Pew figures |
| Negative societal outlook, ages 18 to 29 | 48% | Pew Research Center |
| Ages 30 to 49 | 39% | Pew Research Center |
| Ages 50 and over | 37% | Pew Research Center |
Source: Pew Research Center, how opinions and use of AI differ by age.
6. Non-Users Are Not Waiting to Convert
The half of the population outside this market shows little sign of entering it. Roughly seven in ten adults who do not use chatbots say they are unlikely to start within the next 12 months, including 40% who say they are not at all likely to.
Their stated reasons are lack of interest, privacy concerns, worry about inaccuracy, and not knowing how. Only one of those, the last, is addressed by making products easier to use. Lack of interest in particular is not a barrier that better onboarding removes, and it suggests the remaining population is not a queue of future adopters but a genuinely distinct group. On the smart device side the picture is similarly plateaued: smartwatches reach 37%, smart speakers 35%, smart doorbells 18%, robot vacuums 13%, and smart thermostats 11%. None of these approach the reach of passive AI exposure through search.
| Metric | Value | Source |
|---|---|---|
| Non-users unlikely to start in 12 months | roughly 7 in 10 | Pew Research Center |
| Not at all likely | 40% | Pew Research Center |
| Stated reasons | lack of interest, privacy, inaccuracy, not knowing how | Pew Research Center |
| Smartwatches | 37% | Pew Research Center |
| Smart speakers | 35% | Pew Research Center |
| Smart doorbells | 18% | Pew Research Center |
| Robot vacuums | 13% | Pew Research Center |
| Smart thermostats | 11% | Pew Research Center |
| Non-user share of adults | 51% | Pew Research Center |
Education-sector context sits in our AI in education statistics. Source: Pew Research Center, key findings on views of artificial intelligence.
Summary: Consumer AI by the Numbers
| Metric | Value | Source |
|---|---|---|
| Use AI chatbots | 49% | Pew Research Center |
| Two years earlier | 33% | Pew Research Center |
| Daily users | 24% | Pew Research Center |
| Non-users | 51% | Pew Research Center |
| ChatGPT | 44% | Pew Research Center |
| ChatGPT in 2023 | 18% | Pew Research Center |
| Gemini | 24% | Pew Research Center |
| Copilot | 17% | Pew Research Center |
| Meta AI | 14% | Pew Research Center |
| Searching for information | 42% | Pew Research Center |
| Work tasks, employed adults | 38% | Pew Research Center |
| Medical advice | 20% | Pew Research Center |
| Companionship | 4% | Pew Research Center |
| Read AI search summaries | 60% | Pew Research Center |
| AI advancing too fast | 63% | Pew Research Center |
| Personal information less secure | 71% | Pew Research Center |
| No confidence in government regulation | 67% | Pew Research Center |
| No confidence in corporate responsibility | 59% | Pew Research Center |
| Negative societal impact expected | 40% | Pew Research Center |
| Smartwatches | 37% | Pew Research Center |
| Smart speakers | 35% | Pew Research Center |
| Survey sample | 5,119 US adults | Pew Research Center |
Methodology and Sources
- All figures come from Pew Research Center’s Americans and AI 2026 report, based on a survey of 5,119 US adults conducted 17 to 23 February 2026 through the American Trends Panel (main report, full PDF, detailed appendix tables).
- Chapter-level breakdowns on non-use, definitions, and age differences come from the same report series (why don’t people use chatbots, what do Americans think AI is, how opinions and use differ by age, key findings summary, Pew AI topic hub).
- Data watch: this survey covers United States adults only, so none of these figures describe global consumer AI behaviour and applying them outside the US is not supported. All measures are self-reported, which matters more than usual here because respondents may not recognise AI features embedded in products they use, meaning the 60% figure for AI search summaries likely understates actual exposure. Named-tool percentages overlap heavily and do not form a market share split. The fieldwork window is a single week in February 2026, and consumer AI adoption has been moving several points a year, so the levels will have shifted by the time you read this even though the structural findings should hold. Pew’s American Trends Panel is probability-based, which makes this more rigorous than the opt-in online panels behind most AI adoption statistics in circulation. Rows marked as derived are arithmetic on published figures.
- Last updated: August 2, 2026. We update this roundup quarterly, and the next major refresh is expected when Pew Research Center publishes its next Americans and AI survey.