Introducing Anthropic Interviewer

Source: Anthropic Research Date: 2025-12-04 URL: https://www.anthropic.com/research/anthropic-interviewer

Summary

AI-powered interview tool for qualitative research at scale. Three-stage pipeline: rubric generation, adaptive AI interviewing, human-assisted analysis. Applied to 1,250 interviews across general workforce (1,000), creatives (125), and scientists (125). Results: 86% of workers report time savings; 65% feel AI displacement anxiety; creatives at 97% time savings with 70% experiencing peer stigma; scientists at 79% trust barriers, confining AI to peripheral tasks. All data released publicly.

Implications

This is Anthropic’s qualitative research infrastructure tool — the system that powered the 81,000-person survey. The three-group breakdown tells a richer story than aggregate data: scientists’ trust barriers concentrating AI use in writing/coding rather than hypothesis generation is the most strategically interesting finding. If the highest-credentialed knowledge workers don’t trust AI for core reasoning tasks, the “AI accelerates science” narrative has a ceiling. The peer stigma for creatives (70%) is a social adoption barrier that won’t disappear with capability improvements. Watch the Anthropic Interviewer methodology being adopted for recurring Economic Index qualitative components and for the public data release enabling independent validation of Anthropic’s economic narrative.