About 60% of U.S. adults say they read AI summaries at the top of search engine results, but about 30% say they don’t and 10% aren’t confident.
When Google AI overview appears in search results, people are less likely to click through to traditional results. Another Pew study from last year found that people who saw an AI overview clicked on a traditional search result 8% of the time, compared to 15% if they didn’t see an AI overview. This difference suggests that in some cases, Google AI Overview may replace browsing rather than encourage it.
Deploying and using chatbots
An additional 49% of U.S. adults say they use an AI chatbot. This is up from 33% in 2024, but Pew noted that the wording of the questions has changed from survey to survey. Taken together, the data shows that adoption of AI chatbot tools is on the rise. Approximately 1 in 4 U.S. adults use chatbots every day.
When Pew first measured ChatGPT usage in 2023, 18% of U.S. adults said they had tried ChatGPT. Since then, that percentage has increased to 44%. ChatGPT leads this space, followed by Gemini with 24%, Copilot with 17%, Meta AI with 14%, Grok with 8%, Claude with 6%, and Character.ai with 3%.
Although AI chatbots are primarily used for information retrieval, many adults also report using AI chatbots for work and entertainment. Among all adults, 42% use chatbots to gather information, 38% of employed adults use chatbots for work, and 25% use them for entertainment. About a fifth seek medical advice, diet and fitness information, 13% seek news, 10% seek emotional support and 4% seek companionship.
How do Americans view the impact of AI?
40% of all U.S. adults predict that AI will have a negative impact on society over the next 20 years. In contrast, far fewer people expect positive impacts (16%), 31% think positive and negative impacts are equivalent, and 13% are not sure. Expectations differ depending on whether the focus is on society as a whole or on the life of an individual. While 40% expect AI to have a negative impact on society, a smaller percentage (31%) say it will have a negative impact on their personal lives and 23% expect a positive impact.
About 7 in 10 U.S. adults believe that AI will make their personal information less secure, while 3% say it will make their personal information more secure. A further 10% said there was no significant difference and 16% were unsure.
Almost two-thirds (63%) of adults say AI is progressing too fast, but only 2% say it is progressing too slowly. 19% said they were progressing at about the right pace, and 16% were unsure.
In a previous Pew Research Center survey, more adults expressed concern than excitement about the increasing use of AI in daily life. This pattern continues even as the use of chatbots increases.
Young Adults: Most Used and Deepest Suspicions
Chatbot users come from all walks of life, but younger people are significantly more likely to use chatbot tools than older people. 66% of adults aged 18-29 use an AI chatbot, compared to 61% of adults aged 30-49, 42% of adults aged 50-64, and 23% of adults aged 65 and older. Adults under 50 also say they use chatbot tools more often, especially for information search, work, and entertainment.
Young adults also express deep concerns about AI. Almost half of adults aged 18-29 (48%) predict that AI will have a negative impact on society over the next 20 years, compared to 39% of adults aged 30-49 and 37% of adults aged 50 and over. Younger people are also more likely to say that AI will have a negative impact on their personal lives, with 37% of adults under 30 saying so, compared to 30% of 30-49 year olds and 28% of over 50 year olds.
Regarding creativity, the opinions of young people are fundamentally divided. 25% say chatbots help their creativity, and 20% say these tools hurt their creativity. Older people are more positive about this question.
Half of the country is still not using chatbots
Approximately half of U.S. adults do not use AI chatbots. Most non-users don’t plan to start right away. 67% of people who do not use chatbots say they are unlikely to use a chatbot in the next year, while only 5% say they are very or very likely to use a chatbot.
The survey also asked non-chatbot users why they don’t use a chatbot. The most common reason is lack of interest, with 60% citing this as their main reason. More than half (54%) say they are concerned about how their personal information will be used, and 45% say they don’t trust chatbots to provide accurate information. Some people don’t know how to use it or are afraid of being judged.
There is a big difference in non-use depending on age. About three-quarters of adults over 65 say they don’t use AI chatbots, compared to about one-third of adults under 30.
Regulation and trust in business
Only a small percentage of U.S. adults express strong confidence that the U.S. government will effectively regulate AI. 67% have little confidence in the government’s ability to regulate AI, up from 62% in 2024.
Democrats are now more skeptical than Republicans about the government’s ability to effectively regulate AI. The survey found that 74% of Democrats have little or no confidence in government regulation, compared to 61% of Republicans. In 2024, Republicans were even more skeptical. Pew said distrust among Democrats rose by 20 percentage points over the same period, while distrust among Republicans fell from 70% to 61%.
Many American adults also lack confidence in the role of companies in AI development. Roughly six in 10 respondents say they have little confidence that U.S. companies will develop and use AI responsibly.
why is this important
More and more people are encountering AI-mediated answers before visiting a website, with 6 in 10 U.S. adults reading about AI in search results and 4 in 10 saying they use chatbots to get information. This doesn’t mean that all users will avoid clicking, but it does indicate how much of the search and discovery process is moving to AI-generated surfaces.
What will be particularly interesting going forward is the gap between people’s skepticism and their actions. While people are using these tools, they are also telling researchers that they expect AI to have a negative impact on society, make personal data less secure, or move too quickly. It is an open question whether that gap will widen or narrow as the platform evolves and more people use it.
Looking to the future
Since Pew Research conducted the study in February, Google has expanded AI Mode capabilities, including Personal Intelligence, in nearly 200 countries and territories. We look forward to seeing if this continued expansion changes the relationship between adoption and trust for the rest of the year.
Featured image: Robin Hung/shutterstock
