Anthropic turns 80,508 Claude interviews into a global snapshot of AI hopes and fears

On March 18, 2026, Anthropic used X to point to a new feature page built from a one-week interview project with Claude users. According to Anthropic, 80,508 people from 159 countries and 70 languages participated, making it what the company describes as the largest and most multilingual qualitative study of its kind. The interviews were conducted through an AI interviewer, a Claude-based system designed to gather open-ended responses at scale.

What makes the project notable is not only the size, but the structure of the findings. Anthropic says people did not split cleanly into pro-AI and anti-AI camps. Hope and alarm often showed up in the same interview. The company’s published breakdown of hopes puts professional excellence first at 18.8%, followed by personal transformation at 13.7%, life management at 13.5%, and time freedom at 11.1%. The page also includes user quotes about diagnosis support, business building, learning, and fear of job loss or cognitive dependence.

• Anthropic reports 80,508 completed interviews.
• Respondents came from 159 countries and used 70 languages.
• The study was run with an AI interviewer rather than a static survey form.

That combination matters. Most public conversations about AI still revolve around benchmark wins, policy debates, or small-sample polls. Anthropic’s dataset is different because it asks current users what “AI going well” would actually mean in daily life. The answers are messy in a productive way: people want higher output, less routine work, more economic mobility, and more support with learning or health, while also worrying about dependency, displacement, and social instability. For product teams and policymakers, that kind of mixed signal is arguably more useful than cleanly polarized talking points.

The study is not a complete picture of society because the respondents were self-selected Claude users, not a representative global sample. Even so, it is a substantial primary-source look at real AI usage sentiment at scale. The original X post is here, and Anthropic’s full write-up is here.