Study: Embarrassment Pushes Clients to Choose AI Over Human Advisers

# What the study looked at

# Main finding in one line

# How the experiments worked

Experiment 1 presented scenarios with low or high embarrassment and asked participants whether they would explain the problem to a human or AI adviser. Most chose humans when embarrassment was low.

Experiment 2 manipulated both embarrassment and adviser competence as rated by independent experts. Making AI look more competent pulled choices toward AI. A high embarrassment level also moved people toward AI. If the topic was embarrassing enough or AI was shown as clearly more skilled, participants chose AI.

Experiment 3 asked participants to describe their own embarrassing financial or tech problem to a human or to AI. Responses were scored for direct admissions, detail, and whether participants justified their behavior. Participants offered justifications to humans 33% of the time and to AI 15% of the time.

# Why embarrassment matters

# Practical implications for advisers and firms

• Service mix: Firms can increase uptake among clients with shame-sensitive problems by providing an AI option alongside human advisers. An AI option can act as an entry point for people who otherwise avoid seeking help.
• Communication: When promoting human advisers for technical competence, acknowledge privacy and judgment concerns. Explicit reassurances about nonjudgmental handling may reduce the competence-judgment gap.
• Product design: Position AI for specific, sensitive use cases where candid detail is crucial to diagnosis (for example, clarifying the root cause of debt or malware exposure). Make competence claims clear when you want people to prefer AI on the basis of skill.

# What this does and does not say The study focuses on disclosure behavior under embarrassment and perceived competence. It shows a behavioral trade-off that predicts adviser choice better than competence or embarrassment alone. It does not evaluate long-term outcomes of advice quality or accuracy, nor does it compare real-world advisor performance — it reports on choices and disclosure in controlled scenarios.

# Bottom line Clients weigh two things when choosing between human and AI advisers: how well the adviser can solve the problem and how much the client expects to be judged. Higher embarrassment shifts people toward AI, and higher perceived AI competence does the same. Offering both options can help more clients seek and receive useful advice.