Artificial intelligence (AI) has turn into half of on a regular basis life for a lot of Americans – at work, at school, in health care and past. As AI spreads, the public stays cautious, however considerably open to its potential advantages.
Drawing on 5 years of Pew Research Center surveys, listed here are 13 findings about how Americans use and look at AI, and the place they see promise and threat.
Americans proceed to be cautious of AI’s impression on every day life. Half of U.S. adults say the elevated use of AI in every day life makes them really feel extra involved than excited, in line with a June 2025 survey. Just 10% say they’re extra excited than involved. Another 38% say they’re equally involved and excited.
More Americans are involved right this moment than they have been after we first requested this query in 2021. Back then, 37% mentioned they have been extra involved than excited.
In distinction, concern is decrease in lots of of the 24 other countries we’ve polled about AI.
U.S. adults are usually involved about AI’s impact on creativity and relationships however are extra open to utilizing it for data evaluation. About half of Americans mentioned in the June survey that AI will worsen individuals’s capacity to assume creatively and kind significant relationships with others. Far fewer mentioned AI will make this stuff higher.
However, Americans are typically extra open to AI playing a role in data analysis tasks equivalent to forecasting the climate.
Americans are extra optimistic about AI in medical care however pessimistic about its impression on training and jobs.
Americans have been extra more likely to say AI may have a optimistic (44%) than destructive (19%) impression on medical care in the U.S. over the subsequent 20 years, in line with an August 2024 survey.
Views have been extra destructive for the different areas we requested about. For instance, solely about 1 / 4 mentioned AI would have optimistic impacts on training (24%) and the way individuals do their jobs (23%).
Across the areas we requested about, 16% to 30% mentioned they weren’t positive what impression AI would have.
A majority of teenagers use AI chatbots. Roughly two-thirds of U.S. teenagers ages 13 to 17 (64%) say they ever use an AI chatbot, in line with a fall 2025 survey.
Teens use chatbots for many reasons, however looking for data and getting assist with schoolwork are the most typical ones, out of the makes use of we requested about. More than half of teenagers say they use chatbots for these causes, and one-in-ten teenagers report utilizing a chatbot to assist with all or most of their schoolwork.
About four-in-ten teenagers additionally say they use chatbots to summarize textual content or movies (42%) or to create or edit photos or movies (38%). Fewer say they use chatbots for extra private causes, equivalent to informal dialog (16%) or emotional help or recommendation (12%).
Most teenagers say utilizing chatbots to cheat on schoolwork is frequent at their faculty. About six-in-ten say college students at their faculty use chatbots to cheat a minimum of considerably typically. This contains about a 3rd who say it occurs extraordinarily or fairly often.
A rising share of U.S. employees say a minimum of some of their work is finished with AI. That share has risen from 16% in 2024 to 21% in a September 2025 survey. However, a majority of American employees (65%) nonetheless say they don’t use AI a lot or in any respect of their job.
Public consciousness and utilization of AI proceed to develop. Almost all Americans mentioned that they had heard or learn a minimum of somewhat about AI, together with 47% who mentioned that they had heard lots about it in a June 2025 survey. The latter has elevated 21 share factors since 2022, when 26% mentioned they’d heard lots about AI.
In addition, 31% of Americans mentioned they interact with AI a minimum of a number of instances a day, up from 22% in February 2024.
Younger adults are extra seemingly than older Americans to bear in mind of and use AI.
Around half of adults underneath 50 say they work together with AI about as soon as a day or extra typically. Smaller shares of these 50 and older say the similar, according to the June survey.
We see an identical sample in the shares of employed Americans who say they’ve ever used ChatGPT at work. While 38% of employed 18- to 29-year-olds say they’ve completed this, that drops to 30% of these 30 to 49 and 18% of these 50 and older, in line with a February-March 2025 survey.
When requested about the elevated use of AI in every day life, youthful adults are slightly more excited than older adults. Still, majorities of younger adults are wary of AI’s impression on issues like human creativity or forming significant relationships.
Using AI-assisted gadgets begins early for some children. In a May 2025 survey, we requested mother and father of younger children about their baby’s use of know-how. About one-in-ten informed us that their 5- to 12-year-old makes use of AI chatbots.
And 11% of mother and father with a child 12 or youthful mentioned their baby makes use of a smartwatch, which frequently has AI-assisted options. This share rises to 21% amongst mother and father with an 11- or 12-year-old.
Parents of older youngsters are additionally extra seemingly than these with a youthful baby to say their child makes use of a voice-activated assistant like Siri or Alexa.
11% of mother and father say their baby makes use of a smartwatch, related share with an older child say this about chatbots
% of U.S. mother and father of a baby age 12 or youthful who say that so far as they know, their baby ever makes use of or interacts with the following
* Based on mother and father of a baby age 5 to 12.
Note: Refer to the questionnaire for full query wording and response choices. Parents with multiple baby age 12 or youthful have been requested about one randomly chosen baby and should have youngsters in different age teams. Not all variations proven are statistically important. Those who didn’t reply or gave different responses usually are not proven.
Source: Survey of U.S. mother and father performed May 13-26, 2025.
PEW RESEARCH CENTER
11% of mother and father say their baby makes use of a smartwatch, related share with an older child say this about chatbots
% of U.S. mother and father of a baby age 12 or youthful who say that so far as they know, their baby ever makes use of or interacts with the following
| Voice-activated assistant | Smartwatch | AI chatbot* | |
|---|---|---|---|
| Total | 37 | 11 | 8 |
| Child underneath age 2 | 12 | 7 | N/A |
| Ages 2 to 4 | 27 | 7 | N/A |
| 5 to 7 | 36 | 7 | 3 |
| 8 to 10 | 46 | 13 | 7 |
| 11 to 12 | 53 | 21 | 15 |
* Based on mother and father of a baby age 5 to 12.
Note: Refer to the questionnaire for full query wording and response choices. Parents with multiple baby age 12 or youthful have been requested about one randomly chosen baby and should have youngsters in different age teams. Not all variations proven are statistically important. Those who didn’t reply or gave different responses usually are not proven.
Source: Survey of U.S. mother and father performed May 13-26, 2025.
PEW RESEARCH CENTER
AI specialists are far more enthusiastic about AI than the U.S. public is, however each teams need extra management.
In 2024, we surveyed the American public and experts in the field of AI about their views on AI and its impression on society. Some 56% of AI specialists assume AI may have a really or considerably optimistic impression on the U.S. over the subsequent 20 years, in contrast with simply 17% of the normal public.
However, almost half or more in both groups mentioned they’ve little or no management over AI’s use of their lives, and greater than half need extra management.
Few Americans are getting information from AI chatbots, and people who do have combined views of its accuracy.
Just 9% of U.S. adults say they get information a minimum of typically from AI chatbots primarily based on an August 2025 survey. Most (75%) say they by no means get information this manner.
And seeing inaccurate data is pretty frequent on chatbots. About half of adults who get information from AI chatbots say they a minimum of typically come throughout information there that they assume is inaccurate. That contains 16% who say this occurs typically or extraordinarily typically.
Americans usually tend to say sure teams are higher represented than others in AI design.
Four-in-ten U.S. adults say the individuals designing AI take the experiences and views of White adults into consideration a minimum of considerably properly, according to a 2024 survey. Fewer say the similar about Asian (25%), Black (19%) or Hispanic adults (17%).
The public can be extra more likely to say males’s views are properly thought-about in AI design, in contrast with those that say the similar about ladies’s views.
Still, sizable shares – 37% to 44% – say they’re not sure how properly these teams are represented.
Democrats have much less belief than Republicans do in the U.S. to control AI, in line with a March 2025 survey.
Americans are divided on the quantity of belief they’ve in the nation to control AI use properly. While 44% have a lot or some trust, 47% haven’t an excessive amount of or no belief in any respect.
However, this varies by occasion: 54% of Republicans and Republican-leaning independents have lots or some belief in the U.S. to control AI, in contrast with 36% of Democrats and Democratic leaners.
Note: This is an replace of a submit initially revealed Nov. 21, 2023. Former Associate Director of Research Alec Tyson co-authored the authentic submit.