Aligning Human-AI-Interaction Trust for Mental Health Support: Survey and Position for Multi-Stakeholders

TL;DR

This paper proposes a comprehensive trust framework for AI in mental health support, reviewing current research and outlining a future agenda to align technical and therapeutic trust criteria.

Contribution

It introduces a three-layer trust framework integrating stakeholder perspectives and systematically reviews existing evaluation practices in mental health AI.

Findings

Identifies gaps between NLP metrics and real-world mental health needs.

Highlights the importance of multi-stakeholder perspectives in trust assessment.

Provides a research agenda for developing socio-technically aligned AI.

Abstract

Building trustworthy AI systems for mental health support is a shared priority across stakeholders from multiple disciplines. However, "trustworthy" remains loosely defined and inconsistently operationalized. AI research often focuses on technical criteria (e.g., robustness, explainability, and safety), while therapeutic practitioners emphasize therapeutic fidelity (e.g., appropriateness, empathy, and long-term user outcomes). To bridge the fragmented landscape, we propose a three-layer trust framework, covering human-oriented, AI-oriented, and interaction-oriented trust, integrating the viewpoints of key stakeholders (e.g., practitioners, researchers, regulators). Using this framework, we systematically review existing AI-driven research in mental health domain and examine evaluation practices for ``trustworthy'' ranging from automatic metrics to clinically validated approaches. We highlight critical gaps between what NLP currently measures and what real-world mental health contexts require, and outline a research agenda for building socio-technically aligned and genuinely trustworthy AI for mental health support.