Half of Americans use AI chatbots, but most wish they'd slow down

Americans are using AI more than ever, but they have never been more anxious about it.

A new Pew Research Center survey of 5,119 U.S. adults, conducted in February 2026 and published on June 17, reveals a stark paradox. Roughly half of all Americans (49%) now say they use AI chatbots like ChatGPT, Gemini, or Copilot. That is a massive jump from the 33% reported in summer 2024. Even more telling: about one in four adults now use chatbots on a daily basis, with 4% reporting they use them almost constantly.

Yet, familiarity has not bred comfort. Only 16% of Americans believe AI will have a positive impact on society over the next 20 years, while 40% expect it to be negative.

The young and the skeptical

Conventional tech wisdom says that young people embrace innovation first and worry later. This data completely flips that script.

Adults under 30 are actually the most pessimistic demographic about AI. Only 14% of young adults think the technology will benefit society, and nearly half (48%) anticipate a negative societal impact. While adults under 50 are twice as likely as those 50 and older to use ChatGPT (57% versus 28%), adoption has clearly not translated into optimism.

Public trust in institutions to manage this shift is practically nonexistent. A staggering 67% of Americans say they do not believe the U.S. government will meaningfully regulate AI, and 59% do not trust companies to build it safely. Nearly two-thirds of the country feel the technology is simply advancing too fast.

There is also a modest gender divide in daily habits. While equal shares of men and women use ChatGPT, men are more likely to use chatbots on a daily basis (27% versus 20%) and are more inclined to branch out to alternatives like Microsoft Copilot and Grok.

ChatGPT rules, but AI search is the real daily habit

When it comes to market share, OpenAI is still leaving everyone else in the dust.

ChatGPT is used by 44% of U.S. adults, more than double its reach from 2023. Google’s Gemini sits in second place at 24%, followed by Microsoft Copilot at 17% and Meta AI at 14%. Power-user favorites like Anthropic’s Claude (6%) and Elon Musk’s Grok (8%) remain niche among the broader public, while Character.ai sits at just 3%.

What are people actually doing with these models? Boring, practical things. The top use cases are searching for information (42%) and handling workplace tasks (38% among employed adults). Creative tasks, like making images or videos, hover around 24%, while medical advice and fitness guidance sit at 20%. Despite endless discourse about algorithmic intimacy, just 10% report using chatbots for emotional support, and only 4% use them for companionship.

Meanwhile, the passive infiltration of AI is almost universal. Six in 10 Americans say they routinely read AI summaries at the top of search results. In hardware, 37% wear smartwatches, 35% have smart speakers, and nearly one in five have an AI-enabled doorbell.

Why it matters

This study shows that AI has firmly crossed the chasm from experimental gadget to utility, but users feel like hostages rather than fans.

For workers and developers, the signal is obvious: utility wins over gimmicks. People are coming back every single day because chatbots help their productivity (30% say it helps, compared to just 5% who say it hurts) and keep them informed. You do not need to invent digital soulmates to build a sustainable daily habit; search and work assistance are carrying the entire industry.

For Big Tech, however, this should be a massive wake-up call. Pushing out rapid-fire releases while Silicon Valley celebrates an IPO boom is actively terrifying the public. When nearly 70% of people distrust government oversight and 60% distrust the companies themselves, the industry is operating on a dangerous trust deficit.

We think the disconnect between utility and trust will shape the next era of regulation. Americans will happily use these models to summarize documents, but they will violently push back against autonomous agents or opaque data harvesting. Companies that emphasize security, opt-outs, and unpretentious workplace utility will likely outlast those trying to force artificial companions onto an already nervous public.