The usual story of technology adoption goes like this: a new technology emerges, early adopters adopt it, skeptics follow and eventually join in and the technology becomes normalized. Adoption is fueled by passion—the belief, however naïve in retrospect, that technology is good, that it makes life better, and being an early adopter means being on the right side of something. The story of the internet, smartphone and social media are all versions of this story. People adopted them because they believed in them.
The Pew Research survey of 5,119 US adults It tells a different story set in February 2026. Forty-nine percent of respondents now use AI chatbots—a significant increase in the short term, compared to nearly one-third in 2024. 40 percent say AI will make society worse. Sixty-three percent say AI is moving too fast. 59 percent say they have little or no faith in companies developing AI to do so responsibly. 67 percent say they have very little confidence in the government to regulate it.
What the survey describes is not a technology that people accept because they trust it. It’s a technology that half the country uses despite expecting it to harm the world they live in. This is a condition without clear precedent in the history of mass consumer software adoption.
What past adoption waves looked like
The Internet reached 50% of US household penetration in 2001, at which point it was widely characterized as a transformative force for human communication, commerce, and access to information. The prevailing public mood was optimistic, sometimes excessively so. The smartphone was owned by 50% of US adults in 2012-2013, a period of broad cultural enthusiasm for mobile computing and the connected life it enables. Social media platforms grew in the 2010s on the promise of obvious connection—to connect with people you care about, to find communities organized around shared interests, to have a say in the public conversation. Early criticism of these technologies existed and was sometimes short-sighted, but this was a minority position during an era of rapid adoption. Most people who adopt them do so because they expect them to improve their lives.
This is not to romanticize waves of adoption. Internet optimism led to the dot-com crash. The promise of the smartphone came with a surveillance infrastructure that most early adopters didn’t realize they were taking. Social media’s promise of engagement has added to engagement algorithms that optimize for more outrage from the community. The retrospective picture of each technology is significantly more complex than the enthusiasm that led to its adoption. But the enthusiasm was real, and it was the engine of adoption.
In 2026, AI chatbots are being adopted just as reluctantly. Most Americans who use them expect society to be worse off for it. This is a truly new configuration of the relationship between technology and adoption, and the question it raises is not whether people will continue to use AI—reportedly, they will—but what it means to build an adoption industry driven by nothing but faith.
Why do people use things they don’t trust?
It is not difficult to identify the mechanism for the adoption of trustless artificial intelligence. It’s competitive pressure, or more specifically, the fear of falling short of the standard set by AI. Nearly a quarter of Americans now report using AI chatbots daily; 12% use them several times a day. For this group, AI has become a tool that makes specific tasks faster or easier—writing, researching, coding, summarizing, generating options to consider. A person who uses an AI chatbot to craft an email faster is not making a statement about the impact of AI on society. They make local, rational decisions about productivity. Their expectation that technology will make society worse is a separate belief entered into a separate register.
This separation between what I use and what I think is good is not unique to AI. People use tobacco knowing that it is harmful to their health. People use social media platforms that manipulate their design. People shop at retailers whose labor practices they don’t approve of. Consistency between personal behavior and social values is not a fixed state, but a desire, and the gap between the two is especially wide when the value of consistency is high—which, if not using AI, means being slower, less skilled, or less competitive than the people around you.
What AI has done faster than most previous technologies is the need to compete. Social media adoption was partly social—not being on Facebook or Instagram had social costs in terms of engagement and visibility. Adoption of AI is largely productive—not using it carries efficiency costs in contexts where it is increasingly essential. The sense that you should use it is not driven by peer pressure in a social sense. It is driven by a sense that the alternative falls short of the standard being re-established by people who do not expect to solve wider social issues.
What is insecurity?
The 40% who say AI will make society worse is not monolithic, and understanding what constitutes mistrust is important to understanding what might change it. The Pew data provides some texture. About half of Americans hear from AI chatbots they say that sometimes they come across information that they consider to be inaccurate. Distrust of accuracy is one component—the belief that the technology produces unreliable results—and it coexists with continued use because it is imprecise but still useful for many applications.
The second component is mistrust of management. Sixty-seven percent of Americans have little or no confidence in the government’s ability to regulate AI, and 59 percent have little or no confidence in the companies developing it to do so responsibly. These are not small numbers, and they are not primarily related to the technical capabilities of artificial intelligence. It is about whether the institutions that control the development and application of artificial intelligence can be trusted to make decisions in the public interest. The answer is, for a significant majority of the public, no—and it’s the measure of mistrust that’s most important to understanding the long-term social contract around technology.
The third component is tempo. Sixty-three percent say AI is moving too fast. That’s not the same as saying AI is bad. It is said that the rate of deployment has outstripped the capacity of society – its regulatory frameworks, ethical norms, labor markets, education systems – to adapt. The concern isn’t that AI doesn’t work. It will work faster than the human institutions that must manage its consequences can keep up with it.
What it means to adopt without faith
The companies building the AI economy—OpenAI, Google, Anthropic, Microsoft, Meta—are doing so against a backdrop of public sentiment that is significantly more skeptical than that involved in building previous tech platforms. This is important for reasons beyond public relations. Tech industries that thrive in an environment of public trust tend to benefit from regulatory tolerance, consumer goodwill that embraces early failures, and a cultural narrative that frames industry development as progress. Tech industries that thrive in a climate of public skepticism face the opposite: tighter regulatory scrutiny, less consumer goodwill to embrace failure, and a cultural narrative that frames growth as a manageable challenge.
The current AI industry operates in the latter environment while applying the assumptions of the former. Product introductions, capability announcements, deployment schedules, and investment levels are tailored to a world where the public is, if not enthusiastically, at least widely accepted. Pew data shows that the public is widely used and widely skeptical—a combination that’s more fragile than mere enthusiasm or resistance because it’s held in place by perceived necessity rather than actual purchase. The perceived necessity may evaporate when alternatives appear, the competitive advantage of AI tools narrows, or a sufficiently apparent failure alters the cost-benefit calculus.
A 49% adoption rate is a success story, according to the industry’s own metrics. With half of American adults using a technology’s mainstream application within a few years, it’s an adoption curve that most new technologies fail to achieve. A 2025 YouGov survey found the same pattern a year ago: widespread use, persistent mistrust.
What this adoption rate actually represents—trust or rejection of the technology—is a question the headline number doesn’t answer. Pew does the data, and the answer is instructive. We’ve gotten to the point where people use something they believe in at the rate they used to use something they believed in before. Whether these two conditions produce the same results over time is a question the industry has yet to answer.
