Reach is not consent
Generative features are increasingly bundled into services people already use. Search results are summarized, software drafts text, televisions add assistants, and institutions introduce automated tools into work and education. That expansion can raise usage without showing that people chose the feature, understood the trade-off, or judged it worthwhile.
The distinction matters commercially and politically. Adoption can be driven by defaults, employer requirements, school policy, or the disappearance of a non-AI alternative. Acceptance requires something else: a benefit users recognize, a cost they consider fair, and a credible ability to decline.
The skepticism has measurable foundations
Pew surveyed 5,119 U.S. adults in February 2026. Sixty-three percent said AI was advancing too quickly, compared with 19 percent who said the pace was about right and 2 percent who said it was too slow. Seventy-one percent expected AI to make their personal information less secure, while 3 percent expected more security.
The same report found broad doubts about institutional control: 67 percent had little or no confidence in effective government regulation, and about six in ten were not confident that U.S. companies would develop and use AI responsibly. These findings do not prove that every application is unpopular, but they reject the idea that exposure naturally converts into trust.
A trust campaign cannot substitute for a better bargain
TechCrunch argues that people encounter AI alongside concerns about jobs, creative work, education, privacy, and infrastructure. Data-center developers are adding local concessions as communities scrutinize power, water, jobs, and environmental burdens. Consumer nostalgia for simpler and less mediated tools also reflects a search for agency, not merely resistance to technology.
If the industry's answer is better storytelling, distrust will harden. The stronger response is product and policy evidence: useful services that people can choose, measurable gains shared beyond investors, transparent data practices, credible redress, and local agreements that survive independent scrutiny.
Go to the source
Read the evidence behind this analysis. External links open in a new tab.
TechCrunch — AI was supposed to win people over by now Pew Research Center — Americans and AI 2026
