Deconstructing Depression Stigma: Integrating AI-driven Data Collection and Analysis with Causal Knowledge Graphs

2025-02-09

AI Summary

Depression-related stigma impedes help-seeking and recovery, constituting a critical socio-psychological barrier. Method: This study introduces a novel paradigm integrating LLM-driven conversational data collection with causal knowledge graph construction. A dialogue agent gathered attitudinal data from 1,002 participants; AI-assisted qualitative coding (expert inter-rater reliability κ = 0.92) and causal discovery algorithms were applied to yield interpretable deconstructions of stigmatizing cognition. A cross-dimensional causal knowledge graph of psychological constructs was then built to support both population- and individual-level causal inference. Contribution/Results: This work achieves, for the first time, end-to-end modeling from natural-language dialogues to empirically verifiable psychological mechanisms. It uncovers core transmission pathways of stigma and multilevel psychological antecedents—thereby establishing a methodological foundation and empirical basis for precision interventions and inclusive mental health policy design.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Causal Learning
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Mental-illness stigma is a persistent social problem, hampering both treatment-seeking and recovery. Accordingly, there is a pressing need to understand it more clearly, but analyzing the relevant data is highly labor-intensive. Therefore, we designed a chatbot to engage participants in conversations; coded those conversations qualitatively with AI assistance; and, based on those coding results, built causal knowledge graphs to decode stigma. The results we obtained from 1,002 participants demonstrate that conversation with our chatbot can elicit rich information about people's attitudes toward depression, while our AI-assisted coding was strongly consistent with human-expert coding. Our novel approach combining large language models (LLMs) and causal knowledge graphs uncovered patterns in individual responses and illustrated the interrelationships of psychological constructs in the dataset as a whole. The paper also discusses these findings' implications for HCI researchers in developing digital interventions, decomposing human psychological constructs, and fostering inclusive attitudes.

Problem

• Deconstruct depression stigma using AI.
• Build causal knowledge graphs from data.
• Analyze attitudes toward depression with chatbots.

Innovation

• AI-driven chatbot for data collection
• AI-assisted qualitative conversation coding
• Causal knowledge graphs for stigma analysis