New AI Interviewer Co-Developed with DES Inventor Scales Inner Experience Sampling

For decades, studying subjective inner experience forced a trade-off between depth and scale. Ecological Momentary Assessments (EMA) capture real-time experiences but constrain responses to preset formats. Descriptive Experience Sampling (DES) offers deep expert interviews, but relies on scarce trained interviewers, limiting sample sizes. Now, a team including DES founder Russell T. Hurlburt has built an AI interviewer that operationalizes DES into an explicit, inspectable reasoning architecture — the first LLM-based system grounded in an established phenomenological method.

The AI evaluates each participant message across 11 quality dimensions, maintains a conservative account of established information, selects stage-appropriate interventions, and generates a single non-leading query. Critically, temporal grounding always precedes experiential content. The system was derived from the full corpus of DES transcripts and refined directly with Hurlburt. It runs inside Introscope, an app that delivers randomized beeps, conducts interviews, and lets researchers run studies via shareable links. Pending validation studies, the AI will be freely available for crowdsourced sampling and individual exploration.