AI counselors boost platform activity but shift counselor dynamics in mental health communities

A new arXiv study by Zhang, Zheng, and Yan analyzes the impact of generative AI on two-tiered online mental health communities (OMHCs)—platforms that combine free public Q&A forums with paid private counselor consultations. The researchers leveraged a quasi-natural experiment from a leading OMHC that integrated a genAI-based conversational agent. They found that after AI entry, counselor posting intensity increased significantly, average response length remained unchanged, and per-post social recognition (e.g., likes or upvotes) declined. The AI improved overall responsiveness and expanded patient engagement, enlarging the set of available opportunities for counselors.

Crucially, the effects were heterogeneous: counselors who were intrinsically motivated (driven by altruism) reduced their participation, while economically motivated counselors intensified competitive effort to maintain or grow their downstream paid consultations. This created cross-tier spillovers—inactive public participants saw declines in paid consultations, whereas those who boosted public participation preserved or expanded demand. The authors conclude that in professional tiered platforms, demand expansion and competitive incentives can outweigh intrinsic crowding-out, offering nuanced lessons for AI integration in healthcare.