Forty-nine percent of American adults now use AI chatbots such as ChatGPT, Google Gemini, and Microsoft Copilot, according to a new report from the Pew Research Center. The survey of 5,119 U.S. adults, conducted in February 2026 and released on June 17, paints a vivid portrait of rapid adoption shadowed by deep public skepticism. While nearly half the country has embraced conversational AI, most users harbor serious misgivings about accuracy, privacy, and the very nature of the technology they’re engaging with.

This paradox—rising use alongside low trust—defines the current state of AI chatbots in the United States. For Windows users, many of whom encounter Copilot directly through the taskbar in Windows 11, the findings carry special urgency. Microsoft has staked much of its future on AI integration, but the Pew data suggests that simply putting a chatbot in front of millions of people does not guarantee they’ll trust it.

The State of AI Chatbot Adoption

The 49 percent usage figure marks a sharp increase from earlier surveys. In 2024, Pew found that about 30 percent of Americans had tried a chatbot. The leap to nearly half the adult population in two years underscores how quickly generative AI tools have woven themselves into work, education, and daily life. Among the most popular platforms cited were OpenAI’s ChatGPT, Google’s Gemini, and Microsoft’s Copilot—the latter deeply embedded in Windows 11 and Edge.

Younger adults and those with higher education levels led adoption. Roughly 65 percent of 18- to 29-year-olds reported using AI chatbots, compared to just 28 percent of those 65 and older. College graduates (62 percent) outpaced those with a high school diploma or less (34 percent). Men were slightly more likely than women to be users (52 percent vs. 46 percent).

Yet usage frequency varied widely. Only 22 percent said they used a chatbot daily, while 39 percent turned to one several times a week. The majority—61 percent—used chatbots regularly for tasks like drafting emails, answering questions, or generating creative content. But a substantial minority remained occasional dabblers, testing the waters without fully integrating AI into their routines.

The Trust Chasm: Why Most Users Remain Skeptical

Despite the heady adoption numbers, trust levels were strikingly low. The Pew report revealed that only 29 percent of chatbot users said they had “a lot” or “some” trust in the information these tools provide. A full 71 percent expressed little or no trust. Among non-users, trust was even more anemic, with just 12 percent expressing confidence. Privacy fears loomed large: 68 percent of all adults worried that AI chatbots collect too much personal data, and 63 percent doubted that companies handle that data responsibly.

When asked why they distrust AI chatbots, respondents pointed to three primary concerns: the potential for spreading misinformation (cited by 78 percent), the lack of transparency about how answers are generated (74 percent), and the fear that AI could perpetuate bias or harmful content (66 percent). These anxieties crossed political and demographic lines, revealing a broad societal unease.

“People are using these tools while holding their nose,” said Dr. Emily Norton, a technology researcher at the University of Washington, who reviewed the findings. “They see immediate utility—faster drafting, quick answers—but they don’t internalize the results as reliable. It’s a practical, almost transactional relationship.”

Specific Spotlight: Microsoft Copilot and the Windows Connection

For the Windows community, the Pew data carries distinct implications. Microsoft has aggressively integrated Copilot into Windows 11, placing an icon on the taskbar and threading AI assistance throughout Office apps, Edge, and even system settings. Copilot’s presence is inescapable for many PC users. And yet, trust in Copilot specifically mirrors the general skepticism.

The report found that among users of Microsoft Copilot, 33 percent said they trusted the tool, slightly above the average but still a minority. Microsoft’s enterprise-friendly branding—with its emphasis on data privacy for business accounts—may help. However, consumer users often remain unaware of the safeguards built into the free version. The survey showed that 58 percent of Copilot users did not know how Microsoft handles their conversation data, and 44 percent assumed it was stored and used for training.

These perception gaps highlight a communication challenge for Microsoft. The company has published detailed privacy policies and offers a commercial data protection plan for Copilot, yet the message hasn’t reached the broader public. Windows users on forums and social media frequently express confusion about what Copilot can see, whether it records sessions, and how to delete history.

Privacy Paradox: Using While Worrying

The Pew study quantified what many technology ethicists call the “privacy paradox”—a scenario where people willingly engage with platforms they don’t fully trust. Among AI chatbot users, 72 percent admitted to inputting personal or sensitive information at least occasionally, despite 81 percent saying they’re concerned about data privacy. This contradictory behavior stems from perceived necessity: 54 percent said using chatbots gives them a productivity edge they can’t afford to lose.

For employers and IT administrators managing Windows fleets, this tension is more than academic. Workers may bypass corporate policies to use AI tools that lack proper data protections, creating shadow IT risks. Microsoft’s recent push with Copilot for Microsoft 365 aims to provide a trusted, compliant alternative, but adoption there hinges on convincing users that the enterprise version genuinely differs from the free one.

Demographic Divides in Trust and Usage

Trust, like usage, varies significantly across groups. Hispanic adults (76 percent) and Black adults (73 percent) expressed higher levels of distrust than White adults (64 percent). Women were 8 percentage points more likely than men to say they distrust AI chatbots. Income also played a role: households earning less than $50,000 annually showed greater skepticism (70 percent) than those earning over $100,000 (58 percent).

Age, predictably, split trust levels, but not as dramatically as expected. While 18- to 29-year-olds were more frequent users, their trust (31 percent) wasn’t dramatically higher than that of the 50- to 64-year-old cohort (25 percent). This suggests that digital nativity doesn’t automatically translate into confidence in AI outputs. Instead, heavy use may breed familiarity without necessarily building trust—a dynamic that could shape future product design.

Misinformation Fears: The Core of the Distrust

The overwhelming number-one concern across all demographics was misinformation. Respondents worried that AI chatbots might generate plausible-sounding but false answers, and that people might rely on those answers without verification. Seventy-eight percent of adults—and 74 percent of users—said this fear was a major reason for their distrust. High-profile incidents of AI “hallucinations,” where tools like ChatGPT or Copilot invent facts or citations, have reinforced these anxieties.

The survey included an experimental component: Pew asked a subset of respondents to fact-check a short AI-generated summary of a current event. Only 35 percent correctly identified errors, and among those who did, 58 percent said the experience made them less trusting. Policymakers have cited such findings in calls for mandatory labeling of AI-generated content and stronger platform accountability.

Regulatory and Industry Responses

In the wake of the report, industry leaders and regulators face renewed pressure. The Pew data comes as the U.S. Congress debates the AI Transparency and Accountability Act, which would require chatbot providers to clearly label outputs, disclose training data sources, and offer users easy opt-outs from data collection. Microsoft and Google have publicly supported certain regulatory frameworks, but they also warn that overly broad rules could stifle innovation.

For Microsoft, which employs Copilot as a cornerstone of its Windows ecosystem, the stakes are particularly high. The company recently expanded Copilot’s capabilities with plugins and deeper system control, but user trust remains the linchpin. Without it, the risk is that Copilot becomes a tool people use when forced to—by an employer, for example—rather than one they choose to rely on.

Why Adoption Keeps Growing Anyway

If trust is so low, why does usage climb? The answer lies partly in the utility AI chatbots provide. Respondents who used them for work (38 percent) or school (22 percent) reported tangible time savings—an average of 2.5 hours per week. Creative professionals and coders praised chatbots for sparking ideas and debugging code. Among small business owners, 47 percent said AI tools helped them compete with larger firms.

Another factor is increasing normalization. As chatbots become routine features in search engines, office suites, and operating systems, the barrier to entry lowers. Many Windows 11 users, for instance, don’t actively choose to use Copilot—it’s just there, prompting them to try it. This ambient computing model introduces millions to AI without requiring a deliberate decision.

What Windows Enthusiasts Should Know

For readers of windowsnews.ai, the Pew findings underscore the need to approach AI chatbots with both curiosity and caution. Here are actionable takeaways:

  • Understand data settings in Copilot: Open the Copilot pane, click the three-dot menu, and go to Privacy Settings. You can review and delete your activity, control ad personalization, and turn off data sharing for training (where available).
  • Verify critical information from any chatbot: Use the chatbot as a starting point, not an authority. Cross-check facts with trusted sources, especially for health, financial, or legal queries.
  • Be mindful of context: Copilot in Windows can see active app content if you grant permission—disable this in Settings > Privacy & security > Copilot if you prefer limits.
  • Know the enterprise difference: If your organization uses Microsoft 365, Copilot with commercial data protection does not use your prompts or responses for training. Confirm with your IT admin.

Looking Ahead: Can Trust Catch Up to Usage?

The Pew Research Center’s 2026 report captures a pivotal moment. AI chatbot adoption is no longer a niche phenomenon; it’s approaching majority status. But the trust deficit threatens to limit deeper integration of AI into high-stakes areas like healthcare advice, legal guidance, and personal finance—areas where accuracy and privacy are non-negotiable.

Microsoft, Google, and OpenAI face a delicate balancing act: they must continue innovating while rebuilding public faith through transparency, robust security practices, and consistent messaging. For Windows users, the evolution of Copilot over the next 12 to 18 months will serve as a real-world test case. Will deeper integration and proven reliability close the trust gap, or will it widen as AI surfaces in ever more personal aspects of computing?

One thing is clear from the Pew data: getting half the country to use AI chatbots was the easy part. Convincing them to trust what these tools say—and what they do with their data—is the next great challenge.