Half of U.S. adults use an AI chatbot, and a quarter use one every day, according to a new poll released Wednesday by Pew Research. This is up from 33% of Americans who used an AI chatbot in the summer of 2024. However, a minority (16%) believe that AI will have a positive impact on society.
OpenAI’s ChatGPT is the most popular chatbot in the US, with 44% of Americans reporting using AI tools. When Pew surveyed Americans last year, 34% said they use ChatGPT.
The second most popular chatbot is Google’s Gemini (24%), followed by others: Microsoft Copilot (17%), Meta AI (14%), xAI’s Grok (8%), Anthropic’s Claude (6%), and Character.ai (3%).
51% of U.S. adults said they never use AI chatbots, and 25% said they use them “less than a few times a week.” According to Pew, 24% of respondents said they use chatbots every day, with 8% using them about once a day, 12% several times a day, and 4% “almost always.”
The use of AI chatbots is highly age-dependent. 66% of 18-29 year olds use AI, with the percentage decreasing for each age group: 30-49 (61%), 50-64 (42%), and 65+ (23%).
Pew points out that young Americans are the ones using AI the most, but that doesn’t mean they view it more positively. In fact, 18- to 29-year-olds are the most skeptical, with 48% saying AI will have a negative impact on society over the next 20 years and 37% saying it will have a negative impact on individuals. Comparing this to 30-49 year olds, 39% said they believe AI will have a negative impact on society over the next 20 years, and 30% believe it will have a negative impact on individuals.
People over 50 are the most skeptical of AI, with 37% saying AI will have a negative impact on society over the next 20 years and 28% saying AI will have a negative impact on individuals. And it’s not difficult to understand why such a discrepancy exists. Just look at recent college graduation ceremonies, where young people booed commencement speakers who praised AI, largely because they were told it would take their jobs.
These 50+ people aren’t just staring down a job market where they have to battle AI for entry-level positions, or an application process flooded with fake job postings and AI-generated applications and interviews.
High-income households are most likely to use AI, with 66% of high-income households telling Pew that they use chatbots, 49% of middle-income households saying the same, and just 41% of low-income households saying the same.
What are people using AI for? 42% say they use AI to find information, suggesting that the most common use case for AI is to replace search engines. The second most popular answer was “for tasks at work,” with 38% of employed adults saying they use AI at work.
While many companies are implementing AI tools, and some are strongly encouraging their employees to use them, there are concerns about how AI tools are being used. Sharing proprietary or confidential information with AI systems can be problematic for some companies, and research shows that employees sometimes use unapproved AI tools despite being warned not to do so.
Just 25% of those surveyed by Pew said they use AI for fun and entertainment, and 24% said they use it to create or edit images and videos. 20% say they use AI for medical advice, and 20% also get diet and fitness information. 13% use AI to “get news,” but it’s unclear what exactly that means. 10% say they use AI for “emotional support and advice” and 4% say they use AI for companionship.
It’s no wonder that many people using AI believe it can help improve productivity. 30% of Americans believe AI will help improve productivity, and 5% say it will cause harm. 14% believe it will neither help nor harm. American adults are more skeptical about how AI will impact their well-being and relationships. Only 8% believe it helps their well-being, and 5% say it hurts their well-being. 36% believe it is neither harmful nor effective. Polls on whether it helped their relationship were about the same.
Americans are also skeptical about how AI will affect their country’s information security. 71% think AI will make their personal information less secure, 10% say it won’t make a big difference, and just 3% say AI will make their personal information more secure.
Looking at the racial breakdown of AI chatbot usage, Asian Americans are the most prolific users, with 70% using it a few times a week or less and 47% using it daily. White Americans were the least likely to use AI chatbots overall (46%), with 20% using them daily. Compare this to Hispanics (49% overall, 26% daily) and Black Americans (49% overall, 24% daily).
As for what role the U.S. government might play in the emergence of AI, 67% of U.S. adults say they have little trust in elected leaders to regulate technology. 59% say they are not confident that their companies are developing and deploying AI responsibly.
Politically, majorities of both Democrats and Republicans have little confidence in the government’s ability to effectively regulate AI, but those on the left are the most skeptical. 74% of Democrats do not believe the US government will effectively regulate AI, while 61% of Republicans believe the same. Republicans are more skeptical of the government under President Joe Biden, with 70% of Republicans saying two years ago that they no longer believe the government will be able to effectively regulate AI in 2024.
Regarding reasons for not using AI chatbots, 83% said they were “simply not interested,” 79% were “worried about how their personal information will be used,” 76% said “I don’t think they can get accurate information,” and 55% said they “don’t know how to use them.” 14% think others will judge them for using AI.
According to Pew, the survey was conducted February 17-23 and asked 5,119 U.S. adults about their use of AI.
