I stood in a packed auditorium as graduates booed a commencement speaker who praised AI. You could feel the room tilt from curiosity to alarm in a single breath. That moment compressed a national unease into something public and personal.
On a bus this morning, two strangers traded ChatGPT prompts — and one admitted they use it every day.
I follow the new Pew Research numbers closely because you want the facts that matter. Half of American adults now say they use AI chatbots, a jump from 33% in summer 2024, and only 16% believe AI will have a positive impact on society over the next 20 years. OpenAI’s ChatGPT leads the pack at 44% usage, followed by Google’s Gemini at 24%, Microsoft Copilot 17%, Meta AI 14%, xAI’s Grok 8%, Anthropic’s Claude 6%, and Character.ai 3%.
How many Americans use AI chatbots?
Fifty-one percent of adults say they don’t use chatbots at all. Twenty-five percent use them several times a week or less. And 24% use chatbots daily: 8% about once a day, 12% several times a day, and 4% almost constantly. The Pew poll queried 5,119 adults between Feb. 17–23, 2026.
At a graduation, kids booed because they fear losing jobs to code they didn’t write.
That fear shows up in the age breakdown. Young adults (18–29) are the heaviest users at 66%, then 30–49 at 61%, 50–64 at 42%, and 65+ at 23%. Yet younger users are the most pessimistic: 48% of 18–29-year-olds expect AI will have a negative effect on society; 37% expect it will hurt them personally. By contrast, those 50+ are less skeptical—still wary, but less so.
Do Americans trust AI?
Trust is thin. Seventy-one percent say AI will make their personal information less secure. Sixty-seven percent have little or no confidence that elected leaders can regulate the tech; 59% feel the same about companies that build it. Both parties express doubt, with 74% of Democrats and 61% of Republicans saying the government won’t regulate AI effectively.
I watched an HR manager quietly tell a new hire not to paste confidential contracts into an AI tool.
You can see why many companies worry. Employees are using unauthorized AI tools at work, and that creates risks: leaking proprietary data, inconsistent accuracy, and a headache for compliance teams. Pew finds 38% of employed adults use AI for tasks at work, while 42% use it to search for information — increasingly functioning as a search engine substitute.
At a coffee shop, someone used an AI to rewrite a resume in five minutes.
People use chatbots for many things: entertainment (25%), editing images and videos (24%), medical advice (20%), diet and fitness tips (20%), news (13%), emotional support (10%), and companionship (4%). Among users, 30% say AI helps their productivity, while only 5% say it hurts.
The conversation feels like a crowded subway at rush hour.
A recruiter told me that AI applications have started to clog inboxes with near-perfect but fake resumes.
Income matters. Upper-income adults report 66% chatbot use, middle-income 49%, and lower-income 41%. Racially, Asian Americans appear most active: Pew reports 70% use them several times a week or less and 47% use them daily. White adults report 46% overall use and 20% daily; Hispanic and Black adults report 49% overall use with 26% and 24% daily, respectively.
At my desk, a colleague confessed they won’t try AI because they don’t trust its facts.
Reasons for non-use are straightforward: 83% say they’re not interested, 79% fear how their information would be used, 76% doubt accuracy, 55% don’t know how to use the tools, and 14% fear social judgment for using them.
What do people use AI for?
The most common single use is replacing search—42% use AI to find information. Work tasks account for 38% of reported usage. Creative and practical use cases coexist: image and video editing, health questions, and quick drafting. Yet only a minority sees emotional or social upside: 8% say AI helps their happiness; 36% say it neither helps nor hurts.
A lawyer I spoke with warned that trusting AI answers without verification can be like accepting a confident stranger’s directions.
Seventy-one percent of Americans say AI makes personal data less secure. That anxiety drives calls for stronger oversight. If you follow AI policy debates, you’ve watched regulators and tech firms trade cautious assurances while consumers grow more suspicious.
Public opinion is a cracked mirror reflecting different faces back at us.
At scale, these numbers ask a blunt question about who benefits and who bears risk.
Half the country uses these tools, but only 16% expect a net positive for society. That gap between adoption and optimism is the story: mass usage without mass confidence. You’re seeing a technology that people find useful, yet don’t fully trust with their privacy, careers, or democratic institutions.
I’ve covered policy fights, product launches, and user stories — and the pattern is clear: adoption outpaces explanation, and regulation trails both. If you were advising a leader in government or tech, what would you have them fix first?