March 18, 2026
Survey says… the comments say louder
What 81,000 people want from AI
Anthropic asks the world; commenters roast the site, the vagueness, and the sample
TLDR: Anthropic surveyed 80,508 Claude users across 159 countries about hopes and fears for AI using an in-app interview. Commenters blasted the clunky site, vague categories, and user-only sample, demanding viewpoints from non-users—spotlighting a trust gap between glossy AI storytelling and what skeptics want to see measured.
Anthropic dropped a glossy mega-survey of 80,508 Claude users across 159 countries, asking what they want from AI. The company highlights heartfelt wins (a diagnosis found, small business help) and worries (job loss, “something smarter than us”). But the comment section? On fire. The top trend: UX rage. “Boy is that a terrible website, I tried to find a story and gave up,” snapped one reader, as others begged for a simple link to “the story” under the snowglobe of dots.
Next up, the labeling drama. People mocked vague buckets like “professional excellence,” calling the classifications squishy and unhelpful. Cue the Henry Ford meme: “If I’d asked people what they wanted, they’d say a faster horse,” used to dunk on surface-level takeaways. Then came the propaganda call-out. One commenter summed it up as the usual “this can save the world but also kill us all” PR sandwich.
And the spiciest thread? Sample bias. Since the interviewees were Claude users, critics rolled their eyes at the predictable positivity. They wanted to hear from folks who don’t touch AI at all. Still, some readers praised the Quote Wall and the idea of large-scale interviews—if only the framing were clearer and categories sharper.
Key Points
- •Anthropic interviewed 80,508 Claude users over one week in December across 159 countries and 70 languages.
- •The study used Anthropic Interviewer, a version of Claude, to conduct adaptive conversational interviews.
- •Claude-powered classifiers categorized interviews by wants, fulfillment, fears, occupations (if mentioned), and overall sentiment.
- •Participants reported both tangible benefits from AI and significant concerns, often coexisting within individuals.
- •Anthropic published a Quote Wall to browse responses by region, concerns, and visions, and positions the study as the largest multilingual qualitative effort of its kind.