Only 16% of Americans think AI will help society, even as 44% now use ChatGPT

A new Pew study: 16% positive, ~40% negative

TechCrunch's writeup of a fresh Pew Research study lays out a public that uses AI and distrusts it at the same time. Only 16% of US adults think AI will have a positive impact on society over the next 20 years; about 40% expect a negative one. 67% doubt the government will meaningfully regulate it, 59% distrust companies to build it safely, and nearly two-thirds think development is moving too fast.

The usage numbers cut against the gloom. 44% of US adults now say they've used ChatGPT — more than double the 2023 figure — followed by Gemini (24%), Copilot (17%), Meta AI (14%), Grok (8%), Claude (6%) and Character.ai (3%). The pessimism is sharpest among the young: only 14% of under-30s see a positive impact, the lowest of any age group.

Two readings of the gap: hype machine vs. real demand

The thread divided over whether the 16% means the industry misjudged the public or the public will come around. One side argued the enthusiasm was manufactured:

"You should consider that the industry is just lying to make everyone believe that everyone else is interested. Creating a sense of inevitability. It's the same trick that every ad out there uses, selling you a profitable fantasy as reality." — a commenter on the manufactured-inevitability tactic

A CISO on a break, building an AI startup conceded the bubble from inside: "from the inside the tech bubble it seems that everyone wants to use AI. That is not true for most people." The counter came just as hard:

"They didn't overestimate the interest. ChatGPT is extremely popular and notable… there is huge consumer and business demand for AI so they are justified in making the investments they are." — a commenter pushing back on the overestimation claim

The number that stung HN: Claude in last place

The detail that lit up the thread wasn't the 16% — it was the chatbot ranking. Claude, the model HN reaches for most, came dead last among named tools.

"Claude in 6th place, behind Gemini and Copilot and MetaAI and Grok? No wonder the general public still think AI is junk." — Simon Willison, creator of Datasette

Willison also did the homework, linking the underlying Pew report and noting the question measured whether people ever use each chatbot — exposure, not real usage. The explanation the thread converged on was product-market fit:

"Anthropic has found a great product-market fit among programmers. By comparison, all the rest of the tools non-programmers get exposure to are floundering… 'copilot, fix the spacing issue in this word document' or literally any apple genmoji attempt… just suck." — a commenter on Anthropic's narrow but deep fit

Why coding success doesn't translate to public trust

A deeper strand argued the industry mistakes coding ability for general competence — and that's exactly what the public senses.

"Because it can code they think it can do anything else, but unlike code a lot of other work can't have a bunch of little bugs and mistakes because you can't open up real life and edit it after the fact." — a commenter on why physical-world work resists AI

Others questioned the adoption metrics themselves. One pointed at the opt-out problem: counting 1.8 billion Gmail users as AI users when there's no real way to turn the features off tells you nothing about whether people want them. And a media-skeptic argued the demonization runs the other way — that writers fearful of being "first on the chopping block" tilt coverage negative.

The macro fear under the survey

Underneath the trust gap sits an economic dread the survey only gestures at. One commenter tied the sentiment to the market structure directly:

"50% of the S&P 500 valuation is now directly related to AI as are 40% of new layoffs… Corps are effectively borrowing tons of money in order to build the rope to hang the middle class. The idea that you can lay off the middle class and business will continue as usual is a capitalistic fantasy. Without jobs, people can't afford to buy products built with AI." — a commenter on the layoff-demand paradox

That's the real shape of the 16%. The public isn't rejecting AI because it doesn't work — 44% use ChatGPT. They're skeptical of a future being built faster than they trust, by companies they don't trust, in a market betting on the very layoffs that would empty the demand. The optimists' answer is that adoption converts skeptics in time. The pessimists' answer, from the under-30s most exposed to it, is that they've already seen enough.