---
id: "claim-ai-reduces-impression-management"
type: "claim"
source_timestamps: ["§ When the Topic Is Too Sensitive for a Human Interviewer"]
tags: ["psychology", "sensitive-research", "impression-management"]
related: ["contrarian-ai-better-for-sensitive-topics", "entity-chubbies", "entity-outset", "entity-listen-labs"]
confidence: "high"
testable: true
sources: ["commercial"]
sourceVaultSlug: "hbr-seg-commercial"
originDay: 5
articleStem: "hbr-new-30-ai-scale-customer-research"
sourceUrl: "https://hbr.org/2026/04/how-ai-helps-scale-qualitative-customer-research"
sourceTitle: "How AI Helps Scale Qualitative Customer Research"
---
# AI Interviewers Reduce Social Friction for Sensitive Topics

**Claim.** Respondents are often more comfortable opening up to an AI interviewer than to a human when discussing health conditions, personal insecurities, or sensitive subjects. Because respondents feel they are interacting with a machine, they report less fear of judgment and engage in less **impression management**, leading to more open disclosure. Evidence in the source: a men's-health provider researching **erectile dysfunction** (via [[entity-outset]]) and [[entity-chubbies]] researching **young children** (via [[entity-listen-labs]]).

This claim is the counter-intuitive core surfaced in [[contrarian-ai-better-for-sensitive-topics]].

**Confidence:** high · **Testable:** yes

## Enrichment calibration — empirically supported

This is the vault's **best-substantiated** claim. Peer-reviewed work backs the mechanism: **Lucas et al. (2014)** and **Mell & Gratch (2017)** both found more honest, less self-conscious disclosure to computer/virtual agents than to humans (reduced social-desirability bias); QuestionPro echoes that participants "share more candidly with an AI than with a live person"; Sobowale (2025) argues AI can outperform humans on certain sensitive topics. The named examples remain **case anecdotes**, but the psychological mechanism — reduced impression management with a non-judging agent — is **well established**.