Artificial intelligence is becoming part of everyday life for millions of Americans, but that familiarity is not translating into confidence. A new Pew Research Center survey suggests the public is increasingly exposed to AI tools in work, search, shopping and casual online tasks, yet remains wary of what the technology means for society over the next two decades.
According to the study, just 16% of U.S. adults believe AI will have a positive impact on society over the next 20 years. Roughly 40% expect the technology to do more harm than good, while the rest say its long-term effect will be mixed or are unsure. The findings paint a picture of a country that is using AI more often but still has serious doubts about who controls it, how safely it is being built and whether the government can keep pace.
The tension at the heart of the report is striking: adoption is climbing, but trust is not. Pew found that daily chatbot use has become relatively common, with about one-quarter of Americans saying they use AI chatbots every day. At the same time, a majority of respondents say they do not believe the federal government will meaningfully regulate AI, and many do not trust companies to develop the technology responsibly.
Public enthusiasm lags far behind usage
The Pew findings arrive at a moment when AI products are showing up in more consumer and workplace settings than ever before. Many Americans now encounter AI when they search the web, draft emails, summarize documents, generate images or ask a chatbot for advice. But the report suggests that regular interaction has not eliminated uncertainty. In many cases, it may have intensified it.
Only a small share of Americans say they expect AI’s influence on society to be broadly positive. By contrast, the number who think it will be negative is much larger. That imbalance matters for a technology industry that continues to frame AI as a transformative force for productivity, discovery and economic growth.
The gap between industry excitement and public caution also helps explain why the debate around AI has become so politically and culturally charged. Companies are racing to release new models and features, investors are pouring money into the sector, and AI now touches everything from search and advertising to enterprise software and customer support. Yet the average adult appears far less convinced that the benefits will outweigh the risks.
Most Americans doubt regulation will catch up
One of the clearest warning signs in the study is the public’s skepticism about oversight. Pew says 67% of Americans do not think the U.S. government will do anything meaningful to regulate AI. That is not merely a critique of one policy proposal or one administration. It reflects a broader belief that lawmakers are struggling to understand the technology quickly enough to govern it effectively.
This concern is particularly relevant because AI systems are now being deployed in high-stakes settings, including hiring, education, healthcare, finance, customer service and government services. Each new use case raises new questions about bias, transparency, reliability and accountability. When the public doubts regulation will arrive in time, it can deepen suspicion toward both government and the companies building the systems.
The survey also found that 59% of Americans do not trust companies to develop AI safely. That skepticism is notable because the most powerful AI systems are still largely in the hands of a small group of private firms with enormous computing budgets and access to huge amounts of data. For many people, the issue is not whether AI can be useful. It is whether the firms leading the race will voluntarily place limits on products that can be monetized quickly.
“A majority of Americans say they don’t expect the government to regulate AI in a meaningful way, and many also question whether companies can be trusted to build it safely,” the Pew report indicates, underscoring the scale of public unease.
Young adults are more skeptical than older groups
One of the more surprising findings in the survey is that younger Americans are not the most enthusiastic AI adopters when it comes to attitudes. Instead, people under 30 are the age group most likely to express negative views about AI’s impact on society. Pew says only 14% of respondents in this group think the technology will have a positive effect over the next 20 years.
That detail complicates the common assumption that younger adults, as digital natives, will naturally embrace new AI tools. Younger people may be more comfortable experimenting with technology, but comfort does not necessarily equal confidence. In fact, younger Americans may be the most likely to understand how quickly AI is moving, how often it fails, or how much it could reshape entry-level jobs and information access.
The report also says nearly two-thirds of Americans believe AI development is happening too fast. That concern cuts across age groups and suggests that even among people who benefit from AI-powered products, there is a broad worry that innovation is outpacing public debate, product safety and social adaptation.
Older Americans remain the least likely to use AI
At the other end of the spectrum, people 65 and older are far less likely to use AI in daily life. Pew found that nearly 75% of adults in that age group say they never use AI chatbots. Older adults often encounter AI indirectly through search engines, customer service systems and automated recommendations, but many still avoid direct chatbot use altogether.
That pattern is not just about familiarity. It also reflects preference. Respondents who do not use chatbots commonly say they are simply not interested and have no plans to try them in the future. For a large part of the population, AI remains a tool that feels unnecessary, confusing or untrustworthy rather than indispensable.
ChatGPT dominates, but competitors are gaining ground
Among people who do use AI chatbots, OpenAI’s ChatGPT is by far the most widely adopted. Pew says 44% of U.S. adults now report using ChatGPT, more than double the level seen in 2023. That is an important milestone not only for OpenAI, but for the wider consumer AI market. It suggests that the product category has moved from niche curiosity to mainstream utility for a substantial share of the public.
The next most common chatbot in the survey was Google’s Gemini, used by 24% of adults. Microsoft’s Copilot followed at 17%, while Meta AI reached 14%. Farther behind were xAI’s Grok at 8%, Anthropic’s Claude at 6% and Character.AI at 3%.
The spread indicates that while ChatGPT has become the default AI brand for many Americans, there is still room for competition. Some users encounter AI through their existing tech ecosystem, such as Google or Microsoft products, while others are drawn to companion-style or personality-driven chatbots. The market is still in flux, but the survey suggests no single product has yet completely defined the public experience of AI.
| Metric | Pew finding | What it suggests |
|---|---|---|
| Positive impact on society | 16% | Optimism remains limited despite rising adoption |
| Negative impact on society | About 40% | More Americans expect harm than benefit |
| Government will regulate AI meaningfully | 33% say yes | Most doubt public oversight will keep up |
| Trust companies to develop AI safely | 41% say yes | Majority remain skeptical of industry safeguards |
| Daily chatbot use | About 25% | Chatbots have become part of daily routines for many users |
| ChatGPT usage | 44% | OpenAI remains the dominant consumer AI brand |
| People 65+ who never use chatbots | Nearly 75% | Older adults are far less engaged with the tools |
AI is showing up in how Americans get information
The survey also points to a broader shift in how people consume online information. Six in ten respondents said they regularly read AI-generated summaries on the internet. That figure is especially important because it reflects passive exposure rather than deliberate chatbot use. In other words, many Americans are encountering AI whether or not they actively seek it out.
That trend is especially visible in search. AI-generated summaries are increasingly embedded in major search platforms, sometimes appearing before traditional links or excerpts. For users, the feature can be convenient. For publishers and experts, it raises concerns about traffic, accuracy and attribution. If people rely on machine-generated overviews rather than clicking through to original sources, the information ecosystem may change in ways that are difficult to reverse.
Pew also found smaller numbers of people using AI for fitness or dieting advice. That is consistent with a broader pattern in consumer AI use: people often start with low-stakes tasks, such as research or work support, before experimenting with more personal or health-related questions.
Work and research are the top use cases
Among daily users, the most common purposes are research and job-related tasks. That makes sense given the strengths of today’s chatbots, which are often used to summarize text, brainstorm ideas, draft messages, outline documents or speed up repetitive work. The tools are less about replacing human judgment outright and more about acting as an assistant for information-heavy tasks.
Still, the growing reliance on AI for research is one of the areas where skepticism and usage collide most directly. Users may value speed, but they also know the systems can hallucinate, omit context or present inaccurate information confidently. That gap between convenience and reliability is one reason public opinion remains cautious even as usage rises.
Men report higher use and more enthusiasm than women
Pew also identified a noticeable gender divide in AI attitudes and use. Men are somewhat more likely than women to say they use AI chatbots daily, and they tend to express greater enthusiasm about the technology. Women, by contrast, are more skeptical overall.
According to the survey, daily chatbot use is reported by 27% of men and 20% of women. The report notes that men and women report similar levels of ChatGPT use, but men are more likely to use other services such as Copilot and Grok. That difference suggests that broader interest in AI tools may still be shaped by who is trying multiple products and who is more likely to treat AI as a routine part of work.
The gender gap matters because it mirrors wider disparities seen in technology adoption, workplace AI use and perceptions of digital risk. As with other emerging technologies, the first wave of users may not fully represent the broader population. That can leave companies with a skewed understanding of what consumers want, fear or trust.
Pew’s data shows that men are more likely to integrate chatbots into daily routines, while women remain comparatively cautious about AI’s role in everyday life.
What the data says about the state of AI adoption
The most important takeaway from the survey is not that Americans reject AI outright. It is that they are adopting it selectively, often without embracing the broader narrative around it. People may use chatbots to save time, read summaries, or help with work, while still believing the technology is moving too quickly and may not be adequately controlled.
This pattern is familiar in technology adoption more broadly. Consumers often try new tools for practical reasons before forming a stronger opinion about their long-term effects. But AI differs from many earlier technologies because of the scale of the claims surrounding it. Supporters describe it as a general-purpose engine for productivity and innovation. Critics warn about misinformation, job displacement, surveillance, bias and consolidation of power.
Pew’s findings suggest that the public has not yet resolved those competing narratives. Instead, Americans seem to be living with both realities at once: using AI more often while worrying that its societal consequences could still be negative.
Why this matters for Big Tech
For the companies building AI products, the survey should be read as both a warning and an opportunity. The warning is clear: usage alone will not create trust. The opportunity is that ordinary consumers are already incorporating AI into daily habits, which means the market is still open for products that prove useful, reliable and understandable.
Tech giants are in a particularly strong position because they control the platforms where AI is now being discovered. Search engines, operating systems, productivity suites and social apps are all becoming distribution channels for AI features. That gives major firms a powerful advantage, but it also means they will bear the burden of convincing the public that the technology is not just impressive, but safe and accountable.
Public opinion may also shape regulation in the years ahead. If skepticism remains high and more high-profile failures occur, lawmakers could face greater pressure to impose clear rules on model training, deployment, labeling and consumer protection. If the public continues to view AI as both useful and risky, the result may be slower, more fragmented policy rather than a single comprehensive framework.
The bigger picture: a technology ahead of its social license
Pew’s report captures a central paradox of the current AI moment. The technology is becoming normalized before society has reached a consensus on its value. That creates a mismatch between commercial momentum and public comfort.
Companies are building faster models, adding more features and competing for users in a crowded market. At the same time, many Americans are still asking basic questions: Can AI be trusted? Who checks its outputs? Who is responsible when it gets things wrong? And why does it seem to be arriving everywhere at once?
Those questions are not going away. If anything, they are likely to become more urgent as AI expands into school systems, workplaces, government services and consumer devices. The Pew data suggests that the public is willing to use the technology, but not yet ready to call it a net positive.
That may be the defining reality of AI in 2026: widespread use, limited trust and a great deal of uncertainty about what comes next.
Key takeaways from the Pew survey
- Only 16% of Americans expect AI to have a positive effect on society over the next 20 years.
- About 40% believe AI will have a negative impact.
- 67% do not think the U.S. government will regulate AI meaningfully.
- 59% say they do not trust companies to develop AI safely.
- Roughly one-quarter of adults use AI chatbots daily.
- 44% of U.S. adults say they use ChatGPT, making it the leading chatbot by a wide margin.
- Nearly 75% of Americans aged 65 and older say they never use AI chatbots.
- Six in 10 Americans say they routinely read AI-generated summaries online.
The Pew findings suggest a public that is no longer untouched by AI, but still unconvinced by it. The technology is becoming ordinary in practice, even as many Americans remain uneasy about its long-term consequences. That contradiction may define the next phase of the AI story far more than hype cycles, product launches or market valuations ever will.
