AI-Mediated Qualitative Research
AI-mediated qualitative research uses AI systems to conduct, structure, or analyze interviews and other qualitative data while human researchers define goals, review methods, and interpret findings.
Key points
- Anthropic Interviewer shows the pattern: AI drafts a rubric, conducts adaptive 10-15 minute interviews, and helps analyze transcripts afterward [src-068].
- The benefit is scale. Anthropic used it to run 1,250 interviews that would have been expensive and time-consuming through traditional manual interviewing [src-068].
- The method can ask questions that chat-log analysis cannot answer, such as what people did after receiving AI output, how they felt, and what they want AI to become [src-068].
- Human review remains central: researchers collaborated with the tool to edit plans, interpret transcripts, and decide what patterns mattered [src-068].
- The method has limitations common to qualitative research and AI-mediated self-report: selection bias, demand characteristics, static snapshots, text-only emotional analysis, self-report gaps, researcher interpretation, and limited global generalizability [src-068].
- Anthropic’s personal-guidance study identifies a natural next use: follow-up interviews after guidance conversations could reveal what users did afterward, whether Claude changed their minds, and who they would otherwise have asked [src-073].
- This shows why chat-log analysis and qualitative follow-up complement each other: logs expose patterns at scale, while interviews can measure meaning, context, and downstream outcomes [src-073].
Related entities
Related concepts
- Augmentation-Automation Perception Gap
- AI-Use Stigma
- Human-Agent Collaboration
- AI-Native Organizational Process
- AI Personal Guidance
- High-Stakes AI Guidance