Pew: more chatbot use, but lower trust in AI

A new Pew Research Center report shows a central tension in artificial intelligence adoption: more people are using chatbots and AI-powered devices, but public enthusiasm remains limited. The survey was published on June 17 and helps separate real adoption from social trust.

What happened

Pew Research Center published “Americans and AI 2026,” based on a survey conducted February 17–23, 2026 among U.S. adults. The most visible finding is that 49% of adults say they have used AI chatbots, up from 33% in 2024. Adoption is still growing and is no longer limited to technical users or early enthusiasts.

But the same report shows that usage does not equal trust. Pew reports that 63% of adults believe AI is advancing too quickly. In addition, 40% expect AI's impact on society over the next 20 years to be negative, compared with 16% who expect a positive impact. On regulation, 67% say they have little or no confidence that the U.S. government can regulate AI effectively.

The Verge corroborated the main findings and highlighted the same contradiction: nearly half of Americans now use chatbots, while a majority believes the technology is moving too quickly.

Why it matters

The AI industry often measures progress through models, benchmarks, integrations and enterprise spending. Pew offers another indicator: social acceptance. If people use AI tools for convenience while also distrusting their speed, developers and regulation, adoption can continue with friction.

That friction matters for consumer products, schools, companies and governments. An assistant can help summarize documents or draft text, but trust depends on harder questions: who controls the data, when the system is wrong, how harm is corrected, which tasks should be automated and which decisions require human responsibility.

What changes for users, companies and the AI ecosystem

For users, the report suggests a practical but cautious stance. Many people try chatbots because they solve specific tasks, not necessarily because they trust the broader direction of the industry. For companies, the message is clear: adding AI to a product is not enough. Limits, privacy, traceability, human oversight and correction mechanisms need to be explained.

For the AI ecosystem, the numbers show that regulatory debate is not abstract. Weak confidence in government's ability to regulate AI may create room for state rules, private standards, independent audits and stricter internal policies. It may also pressure companies to document models better and avoid exaggerated promises.

Social and strategic context

From a social perspective, the report confirms that AI adoption is not a simple enthusiasm curve. It is ambivalent adoption: people test, compare and incorporate the technology while also doubting it. That ambivalence can be healthy if it forces more transparent products and less marketing-driven design.

The report should also be read carefully. The survey was conducted in February and published in June; it does not describe a change that happened this week, but a recent public-opinion snapshot released now. Nor does it prove that AI is good or bad for society. What it establishes is a gap between everyday use and institutional trust.

What is still unclear

The report does not show how perceptions will change as new generations use AI more frequently, or which events might increase or reduce public trust. It also does not by itself measure product quality, direct labor effects or educational outcomes. What is confirmed is that chatbot use has grown, social expectations remain cautious, and confidence in regulation and responsible development remains weak.

Sources consulted

Pew Research Center: Read More Verge: Read More by Lía Torres — Social and strategic perspective.