188 Generative AI Statistics for 2026

How many people in the US use generative AI?

About half of US adults or more, and the exact share depends on what is asked. A survey of people who ever use a chatbot, a survey of any use by people of working age and an estimate built from product usage data give three different numbers.

• Daily 24%
• Several times a week or less 25%
• Do not use AI chatbots 51%

• 24% of US adults used AI chatbots daily in February 2026: 8% about once a day, 12% several times a day and 4% almost constantly. A further 25% used them several times a week or less, and 51% did not use them. (Pew Research Center (Americans and AI 2026), Jun 2026)
• 54.6% of US adults aged 18-64 used generative AI in August 2025, up from 44.6% in August 2024, on the Real-Time Population Survey. The 2024 share is the value as revised. (St. Louis Fed (State of Generative AI Adoption 2025), Nov 2025)
• Use of generative AI outside work rose from 36.0% to 48.7% of US adults aged 18-64 between August 2024 and August 2025 in the same survey. (St. Louis Fed (State of Generative AI Adoption 2025), Nov 2025)
• Microsoft estimated that 33.0% of the US working-age population (ages 15-64) used a generative AI product in the second quarter of 2026, which put the US at rank 21 in the report's own list of economies. (Microsoft (Global AI Diffusion Q2 2026), Sep 2026)

Pew asks US adults whether they ever use AI chatbots. The Real-Time Population Survey asks people aged 18-64 about any use of generative AI, for work or outside it; its tracker is a live page and the figures may be revised. Microsoft's figure is an estimate built from product usage data for people aged 15-64. The three are different measures and are not averaged or charted together here.

Who uses generative AI the most?

Young adults, by a wide margin. The age gap shows up in the US, the EU, the UK and Germany, and in the EU survey it is far wider than the gap between men and women.

1. 1 Ages 18-29 66%
2. 2 Ages 30-49 61%
3. 3 Ages 50-64 42%
4. 4 Ages 65+ 23%

• 66% of US adults aged 18-29 said they ever use AI chatbots in February 2026, against 61% of those aged 30-49, 42% of those aged 50-64 and 23% of those aged 65 and over. (Pew Research Center (AI use by age 2026), Jun 2026)
• In 2024 the shares for the same four US age groups were 55%, 41%, 25% and 12%, on a question that was worded differently. (Pew Research Center (AI use by age 2026), Jun 2026)
• In the EU in 2025, 64% of people aged 16-24 had used generative AI tools in the previous three months, against 7% of those aged 65-74. (Eurostat (AI use by individuals 2025), Mar 2026)
• By education, the EU share in 2025 was 49% for people with a high level of education, 26% for a medium level and 22% for little or no formal education. It was 35% for men and 30% for women. (Eurostat (AI use by individuals 2025), Mar 2026)
• In the UK, use of generative AI in the last three months was 75% or over for adults below the age of 45, 18% for those aged 75 to 84 and 8% for those aged 85 or over. It was 69% in London and 51% in North East England (November 2025 to March 2026). (DSIT (Public Engagement Survey 2025/2026), Jul 2026)
• In Germany, 29% of people aged 16 to 29 and 22% of those aged 30 to 49 used AI daily in 2026. The survey asked about AI in general. (Bitkom (AI use by individuals 2026), Apr 2026)
• About 80% of active ChatGPT users in the first months after launch had typically masculine first names, a share that fell to 48% as of June 2025. Nearly half of all messages sent by adults came from users under the age of 26 (OpenAI data, worldwide). (NBER (How People Use ChatGPT), Sep 2025)

The US, EU, UK and German figures come from separate surveys with different questions and age bands, so each is read on its own.

How does generative AI use differ between countries?

Nordic countries report some of the highest use in Europe, and richer economies sit well ahead of poorer ones on Microsoft's estimate. Country figures from different producers answer different questions, so each is read on its own.

1. 1 United Arab Emirates 73.3%
2. 2 Singapore 64.3%
3. 3 Ireland 49.9%
4. 4 France 49.6%
5. 5 Norway 49.4%
6. 6 Spain 45.1%
7. 7 United Kingdom 43.7%
8. 8 Australia 40.7%
9. 9 South Korea 40.6%
10. 10 Canada 38.5%
11. 11 Germany 32.3%
12. 12 India 18.5%
13. 13 China 17.5%

• In the EU in 2025, the highest shares of people using generative AI tools were in Denmark (48%), Estonia and Malta (both 47%) and Finland (46%), and the lowest were in Romania (18%), Italy (20%) and Bulgaria (23%). Outside the EU, Norway was at 56% and Switzerland at 47%. (Eurostat (AI use by individuals 2025), Mar 2026)
• 58% of people aged 16 and over in Germany used AI in 2026; 34% used it at least once a week and 15% daily. The survey asked about AI in general. (Bitkom (AI use by individuals 2026), Apr 2026)
• Weekly use of standalone generative AI systems rose from 34% in 2025 to 51% in 2026 across six countries (Argentina, Denmark, France, Japan, the UK and the US), among online adults surveyed from 14 August to 28 September 2026. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• Daily use of generative AI systems across the same six countries went from 7% in 2024 to 16% in 2025 and 28% in 2026. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• Weekly use in 2026 was 42% in the UK, the lowest in the report, 48% in Finland and 57% in Norway. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• On Microsoft's estimate, 28.8% of the working-age population in the Global North used a generative AI product in June 2026, against 16.2% in the Global South. (Microsoft (Global AI Diffusion Q2 2026), Sep 2026)

Eurostat counts people aged 16-74 who used generative AI tools in the three months before its survey. Bitkom asks people in Germany about AI in general. The Reuters Institute surveys online adults about standalone systems such as ChatGPT, and its six-country average is not a global figure. Microsoft estimates use from product data. A country's share on one measure cannot be set against another country's share on a different one.

Which AI chatbots do people use most?

ChatGPT is the most used chatbot in every survey here, with Gemini second. The surveys ask different things: ever use in the US, weekly use across six countries and use among AI users in Germany.

1. 1 ChatGPT 44%
2. 2 Gemini 24%
3. 3 Copilot 17%
4. 4 Meta AI 14%
5. 5 Grok 8%
6. 6 Claude 6%
7. 7 Character.ai 3%

• The share of US adults who say they ever use ChatGPT went from 18% in 2023 to 23% in 2024 and 34% in 2025, on a question asked of people who had heard of ChatGPT. The 2026 question also asked about other chatbots. (Pew Research Center (Americans and AI 2026), Jun 2026)
• By age, 61% of US adults aged 18-29 say they ever use ChatGPT, against 55% of those aged 30-49, 37% of those aged 50-64 and 19% of those aged 65 and over (February 2026). (Pew Research Center (AI use by age 2026), Jun 2026)
• Across the same four US age groups, from youngest to oldest, Gemini is used by 30%, 34%, 20% and 11%, Copilot by 17%, 22%, 19% and 10%, and Claude by 10%, 9%, 4% and 1%. (Pew Research Center (AI use by age 2026), Jun 2026)
• Across six countries, weekly use of ChatGPT rose from 22% in 2025 to 34% in 2026, Google Gemini from 11% to 24%, Microsoft Copilot from 6% to 11% and Claude from 1% to 7%. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• Among AI users in Germany in 2026, 71% use ChatGPT, 50% Google Gemini, 43% Microsoft Copilot and 35% Meta AI. (Bitkom (AI use by individuals 2026), Apr 2026)

Pew measures ever use among all US adults. The Reuters Institute measures weekly use among online adults in six countries. Bitkom reports shares of people in Germany who already use AI. The three sets of shares are not comparable.

What do people use generative AI for?

Mostly to look things up, get advice and write. Work is one use among several, and health questions and emotional support are on the list too.

1. 1 Search for information 42%
2. 2 Fun or entertainment 25%
3. 3 Create or edit images or videos 24%
4. 4 Medical advice 20%
5. 5 Diet and fitness information 20%
6. 6 News 13%
7. 7 Emotional support or advice 10%
8. 8 Companionship 4%

• 42% of US adults say they use AI chatbots to search for information, and 38% of employed adults use them for tasks at work (February 2026). (Pew Research Center (Americans and AI 2026), Jun 2026)
• 20% of US adults use AI chatbots to get medical advice, 10% for emotional support or advice and 4% for companionship. (Pew Research Center (Americans and AI 2026), Jun 2026)
• Across nine countries, the most common weekly uses of generative AI were researching a topic (24%), getting advice (20%), answering factual questions (18%), recommendations (15%), generating new ideas (14%) and summarising text (13%). (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• 64% of generative AI users across nine countries give a faster and easier answer as a reason for using it, and 56% give its interactivity. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• The share who used generative AI to get information in the last week rose from 11% in 2024 to 24% in 2025 and 38% in 2026 across six countries. Use for news rose from 3% to 6% to 9%. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• In the last year, use of generative AI to create text rose from 14% to 20% across six countries, and use to create images rose from 10% to 19%. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• Among AI users in Germany in 2026, 54% use AI for everyday questions such as cooking or repairs, 50% to write or improve texts, 44% for advice on personal questions, 41% for translation, 40% to summarise longer texts and 34% for health topics. (Bitkom (AI use by individuals 2026), Apr 2026)

Pew's shares are of all US adults. The Reuters Institute surveys online adults in nine countries and uses six of them for trends. Bitkom's shares are of AI users in Germany.

What do the AI companies report about their own users?

Each company publishes its own count on its own definition, so the figures below are not like for like. Some are weekly, some monthly and some undefined, and several are a year old or more.

• Asking 49%
• Doing 40%
• Expressing 11%

• Meta said on its first-quarter 2025 earnings call, on 30 April 2025, that there were "almost a billion monthly actives" using Meta AI across its apps. The figure is more than a year old. (Meta (Q1 2025 earnings call), Apr 2025)
• Microsoft said in October 2025 that it had 900 million monthly active users of AI features across its products, and that its own family of Copilots had passed 150 million monthly active users. Both counts are a year old. (Microsoft (FY26 Q1 earnings call), Oct 2025)
• Microsoft also said in October 2025 that GitHub Copilot had over 26 million users. (Microsoft (FY26 Q1 earnings call), Oct 2025)
• OpenAI said in a report dated 2025 that more than 1 million business customers use its tools and that it serves more than 7 million ChatGPT workplace seats, with ChatGPT Enterprise seats up about 9x year over year. (OpenAI (State of Enterprise AI 2025), 2025)
• Google says its own model APIs processed nearly 32 billion tokens per minute in the third quarter of 2026, 40% more than the previous quarter and more than four times the level of a year earlier. (Alphabet (Q3 2026 earnings remarks), Oct 2026)

Meta counts people using its assistant inside its apps each month. Microsoft counts users of AI features across all its products. OpenAI counts business customers and seats. Google gives a token volume. These are company statements with different scopes and periods, and they are not ranked or added together here.

How many workers use generative AI at work?

Between about a third and a half of workers in the surveys below have used it for their job, and daily use is much lower. Each survey sets its own definition of use and its own time window.

1. 1 Professional, scientific and technical services 65.6%
2. 2 Finance, insurance and real estate 59.2%
3. 3 Educational services 53.0%
4. 4 Transportation and warehousing 21.1%
5. 5 Agriculture 17.5%
6. 6 Accommodation and food services 16.3%

• 30% of US employees used AI in their role a few times a week or more as of May 2026, and 52% a few times a year or more. The question covers AI in general. (Gallup (AI indicator), 2026)
• Daily AI use at work was 42% among US employees in technology, 27% in finance and 22% in professional services in May 2026. (Gallup (AI indicator), 2026)
• 47% of US employees say their organisation has integrated AI technology or tools to improve organisational practices, and 25% say it has communicated a clear plan for integrating AI (May 2026). (Gallup (AI indicator), 2026)
• 40.5% of employed respondents in the US used generative AI for work in August 2026, and 5.3% of total work hours were spent using it, on the Real-Time Population Survey tracker. (GenAI Adoption Tracker (Real-Time Population Survey), 2026)
• Use of generative AI for work among US adults aged 18-64 rose from 33.3% in August 2024 to 37.4% in August 2025 in the same survey series. (St. Louis Fed (State of Generative AI Adoption 2025), Nov 2025)
• A Federal Reserve staff note puts work-related generative AI use at about 41% of US workers in November 2025, with 12% using it daily in the last week, citing the Real-Time Population Survey. (Federal Reserve (FEDS Note on AI adoption), Apr 2026)
• 35.9% of workers in Canada had used generative AI tools in their main job in the previous 12 months, as of March 2026. (Statistics Canada (generative AI use by workers, March 2026), Jul 2026)
• Among Canadian workers who use generative AI tools at work, 31.4% used them daily in March 2026, 38.3% a few times a week, 21.7% a few times a month and 8.7% a few times a year. 8.0% used them for most tasks and 3.6% for almost all tasks. (Statistics Canada (generative AI use by workers, March 2026), Jul 2026)
• 48% of employed people in Germany used AI at work in 2026: 8% daily, 18% at least once a week and 22% less often. (Bitkom (AI use by individuals 2026), Apr 2026)
• 84% of people in Germany who use AI at work do so with their employer's knowledge and 12% without it. 21% of employed people have taken AI training from their employer. (Bitkom (AI use by individuals 2026), Apr 2026)

Gallup and Bitkom ask about AI in general; the Real-Time Population Survey and Statistics Canada ask about generative AI. Gallup reports frequency in the role, Statistics Canada any use in the past 12 months. The shares are not comparable across these surveys.

How many businesses use AI?

Fewer than the worker surveys suggest. Official business surveys count firms, and large firms use AI far more often than small ones. These surveys cover AI in general, and each asks its own question.

View data table

Share of EU enterprises using at least one AI technology in 2025, by size. Large enterprises were more than three times as likely to use AI as small ones.

Period	Value
Small	17%
Medium	30.36%
Large	55.03%

• Overall AI use among US businesses moved between 17% and 20% from December 2025 to May 2026, and between 20% and 23% of businesses expected to be using it in the next six months. The question, in use since November 2025, asks about AI in any business function in the past two weeks. (US Census Bureau (Business Trends and Outlook Survey), May 2026)
• By size, 37% of US firms with at least 250 employees and 32% of firms with 100 to 249 employees used AI in the period ending 3 May 2026, against less than 20% of firms with four or fewer employees. (US Census Bureau (Business Trends and Outlook Survey), May 2026)
• By sector, AI use among US businesses was 39.7% in Information and 33.9% in Finance and Insurance, and around 14% in Retail Trade. (US Census Bureau (Business Trends and Outlook Survey), May 2026)
• A Federal Reserve staff note sets US surveys for late 2025 side by side: about 18% of firms had adopted AI (Census, year-end 2025), while 78% of the labour force worked at firms that had adopted AI and 54% at firms using large language models (Atlanta Fed Survey of Business Uncertainty, November 2025). (Federal Reserve (FEDS Note on AI adoption), Apr 2026)
• In the New York and Northern New Jersey region, 61% of service firms said in August 2026 that they had used AI in their business processes in the past six months, up from 40% in 2025 and 25% in 2024. This is a regional survey, not a national one. (New York Fed (Regional Business Surveys, August 2026), Sep 2026)
• Among manufacturers in the New York and Northern New Jersey region, the share using AI was 51% in 2026, 26% in 2025 and 16% in 2024. (New York Fed (Regional Business Surveys, August 2026), Sep 2026)

The Census Bureau changed its question in November 2025, so only values on the current question are given and no earlier values are shown beside them. A share of firms is a different thing from a share of workers: one large adopting firm counts once in a firm survey but employs many people. The New York Fed figures are for its region only.

What do controlled studies say about generative AI and work?

On set tasks such as writing, support chats and standard coding jobs, people with AI help usually finished faster or did more. Results were weaker or negative on tasks outside what the tools do well and among experienced developers working on their own code.

• Among 5,179 customer support agents at one Fortune 500 software firm, access to a generative AI assistant raised issues resolved per hour by 34% for novice and low-skilled workers and had minimal effect on experienced, highly skilled workers (staggered rollout; 89% of agents were outside the United States, primarily in the Philippines). (Brynjolfsson, Li and Raymond (NBER Working Paper 31161), Nov 2023)
• In a preregistered online experiment with 444 college-educated professionals doing writing tasks from their own occupations, those given ChatGPT took 10 minutes (37%) less than the control group's average of 27 minutes, and evaluator grades rose by 0.45 standard deviations. (Noy and Zhang (working paper), Mar 2023)
• In a preregistered experiment with 758 Boston Consulting Group consultants, those using GPT-4 on 18 tasks inside the AI's capability frontier completed 12.2% more tasks and finished them 25.1% more quickly. (Dell'Acqua et al. (HBS Working Paper 24-013), Sep 2023)
• On a task chosen to be outside that frontier, the consultants using AI were 19% less likely to produce correct solutions, as the paper's abstract words it. (Dell'Acqua et al. (HBS Working Paper 24-013), Sep 2023)
• In a pre-registered field experiment with 776 professionals at Procter & Gamble working on real product innovation problems, individuals working with AI improved solution quality by 0.37 standard deviations over individuals without AI, against 0.24 standard deviations for teams of two without AI and 0.39 for teams with AI. (Dell'Acqua et al. (NBER Working Paper 33641), Apr 2025)
• In a controlled experiment with 95 professional programmers recruited through Upwork, the group with GitHub Copilot completed a set coding task, an HTTP server in JavaScript, 55.8% faster than the control group (May to June 2022). The authors are from Microsoft Research, GitHub and MIT. (Peng et al. (GitHub Copilot experiment), Feb 2023)
• Across three randomised field experiments covering 4,867 software developers at Microsoft, Accenture and an unnamed Fortune 100 company, developers given an AI coding assistant completed 26.08% more tasks (standard error 10.3%). (Cui et al. (three field experiments), Feb 2025)
• In a randomised controlled trial, 16 experienced open-source developers took 19% longer to complete issues when allowed to use early-2025 AI tools. They had expected a 24% speed-up and afterwards still believed AI had sped them up by 20%. (METR (early-2025 developer study), Jul 2025)
• METR's follow-up, started in August 2025, estimated that AI tools changed task time by -18% (confidence interval -38% to +9%) for returning developers and -4% (-15% to +9%) for newly recruited developers, against +19% (+2% to +39%) in the early-2025 study. METR calls the new data an unreliable signal because developers increasingly decline to work without AI. (METR (developer study update), Feb 2026)

Each study has its own setting, sample and task, and the results are not averaged or ranked here. The figures come from the working-paper versions linked, not from later journal versions. Standard deviations and percentages are different units and are left as the authors report them.

Do workers say generative AI saves them time?

Most workers who use it say yes. These are self-reports, though, and the time saved is a small share of all hours worked.

View data table

Share of US employees who use AI at work and say it has improved their productivity, by the number of purposes they use it for (second quarter of 2026). People who use AI for more things are more likely to report a gain.

Period	Value
1-2 purposes	45%
3-4 purposes	66%
5-6 purposes	78%
7 or more	90%

• The share of US work hours spent using generative AI rose from 4.1% in November 2024 to 5.7% in August 2025. Workers' self-reported time savings equal 1.6% of all work hours. (St. Louis Fed (State of Generative AI Adoption 2025), Nov 2025)
• The same authors estimate that generative AI may have increased US labour productivity by up to 1.3% since ChatGPT was introduced. This is the authors' estimate, not a measured figure. (St. Louis Fed (State of Generative AI Adoption 2025), Nov 2025)
• On the tracker for the same survey, time saved through generative AI equalled 2.1% of total US work hours in August 2026, as reported by workers. (GenAI Adoption Tracker (Real-Time Population Survey), 2026)
• 65% of US employees in organisations that have implemented AI say it has had a positive effect on their productivity and efficiency at work, including 17% who say "extremely positive". 14% strongly agree AI has transformed how work gets done in their organisation (May 2026). (Gallup (AI indicator), 2026)
• 77% of US employees who use AI for coding assistance report a positive productivity effect, against 68% of those who use it for writing and editing and 65% for search or research. (Gallup (workplace AI survey, Q2 2026), Jul 2026)
• In OpenAI's own survey of 9,000 workers at almost 100 enterprises that are its customers, users reported saving 40 to 60 minutes per active day with AI, and 75% said they could complete tasks they previously could not perform. (OpenAI (State of Enterprise AI 2025), 2025)
• On average across 32 countries, 62% of workers say AI saved them time at work in the last 12 months: 68% of Gen Z, 65% of Millennials, 57% of Gen X and 46% of Baby Boomers. (Ipsos (AI Monitor 2026), Jun 2026)
• In Danish workplaces that combine encouragement, enterprise chatbots and training, 19% of workers in exposed occupations report saving more than one hour a day. 85% of chatbot users say they move the time saved to other job tasks. (Humlum and Vestergaard (NBER Working Paper 33777), Mar 2026)
• Among UK businesses that had adopted AI, 50% reported an increase in productivity. On operating costs, 24% reported reductions, 64% no immediate financial effect and 9% an increase; the report's summary does not state the base for the cost shares (survey 7 January to 21 April 2026). (UK government (AI adoption research 2026), Oct 2026)

Everything in this section is what workers or firms say about themselves, not a measured outcome. OpenAI's survey covers its own enterprise customers. The Real-Time Population Survey, Gallup and Ipsos figures come from different samples and questions.

How many students use generative AI?

Most do. A majority of US teenagers and nearly all UK undergraduates in the surveys here report using AI, much of it for schoolwork, while many parents and college teachers think it harms critical thinking.

View data table

Share of full-time UK undergraduates using AI in at least one way, by survey edition. The 2026 edition polled 1,054 students in December 2025. All three years are printed in the 2026 report.

Period	Value
2024	66%
2025	92%
2026	95%

• 57% of US teens aged 13 to 17 have used AI chatbots to search for information, 54% to get help with schoolwork, 47% for fun or entertainment, 16% for casual conversation and 12% for emotional support or advice. (Pew Research Center (How Teens Use and View AI), Feb 2026)
• 10% of US teens say they do all or most of their schoolwork with the help of AI chatbots, 21% some and 23% a little; 45% have not used them this way. (Pew Research Center (How Teens Use and View AI), Feb 2026)
• 59% of US teens say students at their school use AI chatbots to cheat at least somewhat often, including 34% who say extremely or very often. 14% say rarely or never and 15% are not sure. The question is about what teens see at their school. (Pew Research Center (How Teens Use and View AI), Feb 2026)
• In 2025, 54% of US K-12 students and 53% of English language arts, maths and science teachers said they used AI for school, increases of more than 15 percentage points on surveys one to two years earlier. (RAND (AI Use in Schools, RR-A4180-1), Sep 2025)
• The share of US high school students who report using generative AI tools for schoolwork rose from 79% in January 2025 to 84% in May 2025, and 69% reported using ChatGPT for school assignments and homework in May 2025 (College Board's own surveys). (College Board (student AI research), Oct 2025)
• 86% of US children aged 9 to 17 use or interact with AI: 81% of 9- to 12-year-olds, 89% of 13- to 15-year-olds and 92% of 16- to 17-year-olds. 24% do so daily (survey run 18 to 26 March 2026). (Common Sense Media (AI Use by Tweens and Teens 2026), 2026)
• 75% of US children aged 9 to 17 have used AI-generated search summaries and 67% have used AI chatbots. ChatGPT is the AI used most often for 40%. (Common Sense Media (AI Use by Tweens and Teens 2026), 2026)
• Among US children aged 9 to 17 who use AI, 85% use it for help with schoolwork or homework and 89% for fun or entertainment. 48% use it for schoolwork at least once a week and 21% every day: 12% of 9- to 12-year-olds, 24% of 13- to 15-year-olds and 30% of 16- to 17-year-olds. (Common Sense Media (AI Use by Tweens and Teens 2026), 2026)
• In the EU in 2025, using generative AI for formal education such as school or university was most common among people aged 16 to 24, at 62%. Across the working-age groups from 25 to 64, over 50% used it for work-related purposes. Eurostat's text does not print the base for these shares. (Eurostat (AI use by individuals 2025), Mar 2026)

Pew counts US teens aged 13 to 17 who use AI chatbots. Common Sense Media counts children aged 9 to 17 who use or interact with any AI, which includes AI search summaries, so it is a broader measure. RAND and the College Board use their own panels. These shares are not comparable with each other.

How much money is going into generative AI?

Private investment in AI rose sharply in 2025 on Stanford's count, and most of it went to US companies. Estimates of what businesses spend come from an investor's own model and are labelled that way.

1. 1 United States $285.9bn
2. 2 China $12.4bn
3. 3 United Kingdom $5.9bn

• Global private investment in AI reached $344.7 billion in 2025, a 127.5% increase on 2024. (Stanford HAI (AI Index 2026), 2026)
• Of US private AI investment in 2025, $163.6 billion was generative AI-related. (Stanford HAI (AI Index 2026), 2026)
• Menlo Ventures, a venture capital firm, estimates that US enterprises spent $37 billion on generative AI in 2025, up from $11.5 billion in 2024, with $19 billion going to applications. This is the firm's own estimate. (Menlo Ventures (State of Generative AI in the Enterprise 2025), Dec 2025)
• Menlo Ventures, whose report draws on its own market model and a survey of 495 US enterprise AI decision-makers, says 76% of enterprise AI use cases were bought and not built internally in 2025, against 53% in 2024, and that 50% of developers use AI coding tools daily. (Menlo Ventures (State of Generative AI in the Enterprise 2025), Dec 2025)
• Menlo Ventures estimates that Anthropic earned 40% of enterprise spending on large language model APIs in 2025 (24% in 2024 and 12% in 2023), OpenAI 27% (50% in 2023) and Google 21% (7% in 2023). Menlo Ventures is an investor in Anthropic. (Menlo Ventures (State of Generative AI in the Enterprise 2025), Dec 2025)

Stanford's figures are private investment in AI companies. Menlo Ventures' figures are a venture firm's estimate of enterprise spending and of model providers' shares of API spending. The two measure different things and are not added together.

Do people trust generative AI?

Not much. More Americans expect AI to harm society than to help it, and trust in AI answers is far higher among people who use the tools than among those who do not.

• Negative 40%
• Equally positive and negative 31%
• Positive 16%
• Not sure 13%

• 63% of US adults say AI is advancing too quickly and 2% say too slowly (February 2026). (Pew Research Center (Americans and AI 2026), Jun 2026)
• 71% of US adults think the increased use of AI will make their personal information less secure, and 3% think it will make it more secure. (Pew Research Center (Americans and AI 2026), Jun 2026)
• 67% of US adults have not too much or no confidence in the US government to regulate AI effectively, up from 62% in 2024. 59% have not too much or no confidence in US companies to develop and use AI responsibly. (Pew Research Center (Americans and AI 2026), Jun 2026)
• Among people across nine countries who used AI for the task, 75% trust its answers to factual questions, 76% its advice and 72% its news. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• Among all respondents across nine countries, ChatGPT is trusted by 37%, Gemini by 27%, Copilot by 18% and Claude by 13%. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• 12% of people across nine countries are comfortable with news made entirely by AI, and 19% with AI-made news that has some human oversight. (Reuters Institute (Generative AI and News Report 2026), Oct 2026)
• 20% of people trust news from AI chatbots, on the average of the Digital News Report's markets: 44% of chatbot users and 17% of non-users. That compares with 37% who trust most news most of the time. (Reuters Institute (Digital News Report 2026), Jun 2026)
• Weekly use of AI chatbots for news rose from 7% to 10% in a year across the Digital News Report's markets. 17% of the youngest age group use chatbots for news against 5% of the oldest. (Reuters Institute (Digital News Report 2026), Jun 2026)
• 4% of respondents in 27 markets always or often click through to news sources from AI, compared with 19% from search and 17% from social media. (Reuters Institute (Digital News Report 2026), Jun 2026)
• On average across 32 countries, 54% say products and services using AI have already changed their lives in the past three to five years, and 66% expect AI to have a greater impact in the next three to five. (Ipsos (AI Monitor 2026), Jun 2026)
• On average across 32 countries, 49% think the potential benefits of AI outweigh its environmental costs and 39% disagree. 46% would trust a generative AI tool less if its answers were influenced by advertisers on the platform. (Ipsos (AI Monitor 2026), Jun 2026)
• In a 47-country survey of more than 48,000 people, 66% said they use AI regularly, 46% were willing to trust AI systems, 83% expected a wide range of benefits and 70% believed AI regulation is needed (data collected November 2024 to mid-January 2025). (University of Melbourne and KPMG (global study 2025), May 2025)
• The same 47-country study reports that 66% rely on AI output without evaluating its accuracy and 56% are making mistakes in their work due to AI. The summary page does not state the base for these two shares. (University of Melbourne and KPMG (global study 2025), May 2025)
• 74% of UK adults named inaccurate information as a risk of AI use, 69% privacy and security, 64% the spread of false or misleading information, 63% loss of jobs due to AI, 62% intentionally misleading images or videos and 57% a negative effect on people's ability to think creatively or critically. 3% had no concerns at all (November 2025 to March 2026). (DSIT (Public Engagement Survey 2025/2026), Jul 2026)

Pew surveys US adults. The Reuters Institute reports cover online adults in their own sets of countries, and trust among people who used AI for a task is a different measure from trust among everyone. The Ipsos and KPMG studies are multi-country averages of AI in general.

Is generative AI replacing jobs?

Not on a large scale in the surveys here: most firms using AI report no change in headcount. Exposure estimates say which jobs could be affected, not how many will go, and counts of announced cuts depend on the reason employers give.

• Cuts where employers gave AI as the reason were approximately 20% of the 617,189 job cuts announced by US-based employers in January to October 2026, on Challenger, Gray & Christmas's count of announcements. (Challenger, Gray & Christmas (October 2026 report), Nov 2026)
• Employers gave AI as the reason for 1,265 announced job cuts in October 2026, and for 193,226 since the firm began tracking that reason in 2023. (Challenger, Gray & Christmas (October 2026 report), Nov 2026)
• In the New York and Northern New Jersey region, 4% of service firms said in August 2026 that they had laid off workers in response to AI over the past six months, against 1% a year earlier, and no manufacturers did. This is a regional survey. (New York Fed (Regional Business Surveys, August 2026), Sep 2026)
• About 15% of service firms in the New York and Northern New Jersey region said they had hired fewer workers than they otherwise would have because of AI (12% a year earlier), and about 13% said they had hired more workers due to AI. (New York Fed (Regional Business Surveys, August 2026), Sep 2026)
• Among UK businesses using AI to improve business operations, 63% reported no change in headcount, 1% an increase and 6% a decrease. Around 10% of UK businesses report automating or replacing roles with AI technologies (June 2026). (ONS (AI in UK businesses 2023 to 2026), Jul 2026)
• Globally, one in four workers are in an occupation with some exposure to generative AI, and 3.3% of global employment falls into the highest exposure category, on the ILO's 2025 estimate. (ILO (Generative AI and Jobs 2025), May 2025)
• On the same ILO estimate, 4.7% of female employment and 2.4% of male employment worldwide is in the highest exposure category (9.6% against 3.5% in high-income countries). Employment with some exposure is 34% in high-income countries and 11% in low-income countries. (ILO (Generative AI and Jobs 2025), May 2025)
• The IMF said in January 2024 that almost 40% of global employment is exposed to AI. In advanced economies about 60% of jobs may be affected, and exposure is expected to be 40% in emerging markets and 26% in low-income countries. (IMF (blog on AI and jobs), Jan 2024)
• Employers surveyed by the World Economic Forum expect 170 million jobs to be created and 92 million displaced worldwide by 2030, a net increase of 78 million jobs or 7% of today's employment. This forecast was made in January 2025 and covers all trends combined, not AI alone. (World Economic Forum (Future of Jobs Report 2025), Jan 2025)
• 86% of employers surveyed by the World Economic Forum expect AI and information processing technologies to transform their business by 2030. Half plan to re-orient their business in response to AI, two-thirds plan to hire people with specific AI skills and 40% anticipate reducing their workforce where AI can automate tasks. (World Economic Forum (Future of Jobs Report 2025), Jan 2025)

Challenger, Gray & Christmas counts announced job cuts and the reason the employer stated; an announcement is not a completed layoff. The ILO and IMF figures are exposure estimates made with different methods, so they are not comparable with each other and are not forecasts of job losses. The New York Fed survey is regional, and the ONS figures are for the UK.

What rules apply to generative AI?

The EU has a dated legal timetable under its AI Act, and the US Copyright Office has set out its position on AI output in a report published in parts. The dates below are as printed by each body.

• The EU AI Act, Regulation (EU) 2024/1689, entered into force on 1 August 2024 and became applicable on 2 August 2026, with some exceptions. (European Commission (AI Act page), Oct 2026)
• Under the AI Act, prohibited AI practices and AI literacy obligations entered into application from 2 February 2025, and the governance rules and the obligations for general-purpose AI models became applicable on 2 August 2025. (European Commission (AI Act page), Oct 2026)
• The AI Act's rules for high-risk AI systems embedded into regulated products (Annex I) have an extended transition period until 2 August 2028, and the rules for high-risk use cases in certain sensitive areas (Annex III) have been extended to 2 December 2027, following the political agreement on the AI Omnibus proposal. (European Commission (AI Act page), Oct 2026)
• The AI Act prohibits nine practices. Prohibitions 1 to 8 became effective in February 2025, and prohibition 9, introduced as part of the AI Omnibus, comes into effect in December 2026. (European Commission (AI Act page), Oct 2026)
• The US Copyright Office received over 10,000 comments on its 2023 notice of inquiry on AI and copyright. It published Part 1 of its report, on digital replicas, on 31 July 2024, Part 2, on copyrightability, on 29 January 2025 and a pre-publication version of Part 3, on generative AI training, on 9 May 2025. (US Copyright Office (AI initiative page), May 2025)
• The US Copyright Office concluded in January 2025 that prompts alone do not give users enough control to be the authors of generative AI output. It said it had registered hundreds of works that incorporate AI-generated material, with registration covering the human author's contribution. (US Copyright Office (Part 2: Copyrightability), Jan 2025)

These are dates and positions as published, with no view on their effect. The Commission's page was last updated in October 2026 and reflects the later dates agreed under the AI Omnibus. The Copyright Office's conclusion is its own reading of US law, not a court ruling.

Methodology and sources

Every figure on this page comes from a primary source that was opened and checked: official statistics (US Census Bureau, Eurostat, ONS, Statistics Canada, the UK science department), survey organisations (Pew Research Center, Gallup, the Reuters Institute, Ipsos, Bitkom), central bank research, working papers, company filings and earnings remarks, and regulators' own pages. Measures are kept apart. US adoption is given three ways (ever use of chatbots, any use of generative AI at ages 18-64, and a usage-based estimate) and the three are never merged, averaged or charted together. Business surveys count firms and cover AI in general. Company user counts are each company's own statement with its date and are not compared. Controlled studies are described with their setting, sample and task, using the working-paper version linked. Where a series was revised, the current value is used, and values from before and after a change of question are not shown as one series. Forecasts and estimates are labelled with who made them and when. Regional surveys are labelled as regional. Figures that could not be confirmed on the producer's own page were left out, and a global energy figure is omitted for that reason. The page is reviewed monthly.

• Alphabet (Q3 2026 earnings remarks) · Oct 2026
• Anthropic (Economic Index, January 2026) · Jan 2026
• Anthropic (Economic Index, March 2026) · Mar 2026
• Bitkom (AI use by companies 2026) · Sep 2026
• Bitkom (AI use by individuals 2026) · Apr 2026
• Brynjolfsson, Li and Raymond (NBER Working Paper 31161) · Nov 2023
• Challenger, Gray & Christmas (October 2026 report) · Nov 2026
• College Board (faculty AI research) · Feb 2026
• College Board (student AI research) · Oct 2025
• Common Sense Media (AI Use by Tweens and Teens 2026) · 2026
• Cui et al. (three field experiments) · Feb 2025
• Dell'Acqua et al. (HBS Working Paper 24-013) · Sep 2023
• Dell'Acqua et al. (NBER Working Paper 33641) · Apr 2025
• Digital Education Council (Global Survey 2026) · 2026
• DSIT (Public Engagement Survey 2025/2026) · Jul 2026
• European Commission (AI Act page) · Oct 2026
• Eurostat (AI use by individuals 2025) · Mar 2026
• Eurostat (AI use in enterprises 2025) · Dec 2025
• FBI IC3 (2025 Annual Report) · 2026
• Federal Reserve (FEDS Note on AI adoption) · Apr 2026
• Gallup (AI indicator) · 2026
• Gallup (workplace AI survey, Q2 2026) · Jul 2026
• GenAI Adoption Tracker (Real-Time Population Survey) · 2026
• HEPI and Kortext (Student Generative AI Survey 2026) · Mar 2026
• Humlum and Vestergaard (NBER Working Paper 33777) · Mar 2026
• ILO (Generative AI and Jobs 2025) · May 2025
• IMF (blog on AI and jobs) · Jan 2024
• Internet Watch Foundation (2025 data) · Mar 2026
• Ipsos (AI Monitor 2026) · Jun 2026
• Menlo Ventures (State of Generative AI in the Enterprise 2025) · Dec 2025
• Meta (Q1 2025 earnings call) · Apr 2025
• METR (developer study update) · Feb 2026
• METR (early-2025 developer study) · Jul 2025
• Microsoft (FY26 Q1 earnings call) · Oct 2025
• Microsoft (Global AI Diffusion Q2 2026) · Sep 2026
• NBER (How People Use ChatGPT) · Sep 2025
• New York Fed (Regional Business Surveys, August 2026) · Sep 2026
• Noy and Zhang (working paper) · Mar 2023
• ONS (AI in UK businesses 2023 to 2026) · Jul 2026
• OpenAI (How Organizations Use AI) · Aug 2026
• OpenAI (State of Enterprise AI 2025) · 2025
• Peng et al. (GitHub Copilot experiment) · Feb 2023
• Pew Research Center (AI use by age 2026) · Jun 2026
• Pew Research Center (Americans and AI 2026) · Jun 2026
• Pew Research Center (How Teens Use and View AI) · Feb 2026
• RAND (AI Use in Schools, RR-A4180-1) · Sep 2025
• Reuters Institute (Digital News Report 2026) · Jun 2026
• Reuters Institute (Generative AI and News Report 2026) · Oct 2026
• St. Louis Fed (State of Generative AI Adoption 2025) · Nov 2025
• Stanford HAI (AI Index 2026) · 2026
• Statistics Canada (generative AI use by workers, March 2026) · Jul 2026
• UK government (AI adoption research 2026) · Oct 2026
• University of Melbourne and KPMG (global study 2025) · May 2025
• US Census Bureau (Business Trends and Outlook Survey) · May 2026
• US Copyright Office (AI initiative page) · May 2025
• US Copyright Office (Part 2: Copyrightability) · Jan 2025
• US Department of Energy (data centre report release) · Dec 2024
• World Economic Forum (Future of Jobs Report 2025) · Jan 2025