gizmodo.com web signal

Pew: Just 16% of Americans see AI as good for society

TL;DR

  • A Pew survey of 5,119 U.S. adults conducted February 17-23, 2026 found only 16% believe AI will have a positive impact on society.
  • Half of American adults now use AI chatbots and a quarter use them daily, up from 33% adoption in summer 2024.
  • ChatGPT leads at 44% usage, followed by Gemini at 24%, Copilot 17%, Meta AI 14%, Grok 8%, and Claude 6%.

Half of American adults are now using AI chatbots, a quarter of them every day, and only 16% of the country believes any of it will be good for society. That is the most useful single sentence to come out of Pew Research's latest survey, as reported by Gizmodo, because it captures the strange shape of consumer AI in 2026: heavy use, low trust, no patience for the politicians who are supposed to sort it out.

The numbers underneath are worth lingering on. Pew surveyed 5,119 U.S. adults between February 17 and 23, 2026. Chatbot use is up to 50% of adults from 33% in summer 2024. ChatGPT is the runaway leader at 44%, with Gemini at 24%, Copilot at 17%, Meta AI at 14%, Grok at 8%, and Claude at 6%. The age pattern is the interesting one. Among 18- to 29-year-olds, 66% use AI, and yet 48% of that same cohort say AI will negatively impact society. The people using these tools the most are also the most convinced they are bad for the country.

The trust picture is similarly grim. 71% of respondents told Pew that AI will make their personal information less secure, against just 3% who expect it to improve security. 67% have little to no confidence in elected leaders to regulate the technology. Among the half of Americans who don't use chatbots, 83% say they simply aren't interested, and 76% don't trust the accuracy.

The honest caveat is that polling on AI attitudes is famously sensitive to question wording, and the reporting doesn't tell us how 'positive impact' was framed or whether daily users feel differently from occasional ones. Both would change the read meaningfully. What is harder to dispute is the gap between behavior and belief, and the political vacuum it implies. If two-thirds of adults don't believe Washington will regulate AI competently, the regulatory action is going to come from somewhere else, most likely state attorneys general, plaintiffs' lawyers, and the occasional viral incident. For builders, the strategic opening is trust-as-a-feature: the vendor that can credibly answer the security and accuracy concerns Pew measured has a lot of room to grow into.

Shared on Bluesky by 2 AI experts