Gen AI for metal help-seeking - Survey Data

The deployment of artificial intelligence (AI) in mental health services has increased substantially, yet the specific AI characteristics that influence mental health help-seeking behaviors remain unclear. Drawing on dual-process theory and the Heuristic-Systematic Model (HSM), this study examines the mechanisms by which various characteristics of generative AI influence the intention to seek help. Analyzing data from 619 generative AI users in China, the findings indicate that systematic cues, including perceived intelligence and AI compatibility, along with heuristic cues, including anthropomorphism and likability, positively affect help-seeking intentions through both cognitive and emotional trust pathways. Notably, emotional trust exerts a stronger influence on help-seeking intentions than cognitive trust and can enhance cognitive trust, introducing a potential bias effect. Moreover, perceived privacy risks negatively moderate the mediating effects of cognitive and emotional trust between heuristic and systematic cues and help-seeking intention. Additionally, while users' trust positively correlates with help-seeking behavior, the perceived threat posed by AI can negatively moderate this relationship. This study not only enriches the theoretical discourse surrounding the attributes of generative AI and their impact on mental health help-seeking but also offers practical insights for the design and utilization of AI technologies to enhance mental health support.

人工智能（AI）在心理健康服务领域的部署已大幅增加，但目前仍不明晰哪些具体的AI特征会影响心理健康求助行为。本研究基于双加工理论（dual-process theory）与启发式-系统式模型（Heuristic-Systematic Model, HSM），探究生成式AI（generative AI）的各类特征影响个体求助意向的作用机制。通过分析中国619名生成式AI用户的调研数据，本研究发现：包括感知智能与AI适配性在内的系统式线索，以及包括拟人化与亲和力在内的启发式线索，均通过认知信任与情感信任两条路径对求助意向产生正向影响。值得注意的是，情感信任对求助意向的影响强度高于认知信任，且情感信任可强化认知信任，进而产生潜在的偏差效应。此外，感知隐私风险会对启发式、系统式线索与求助意向之间的认知、情感信任中介效应产生负向调节作用。另外，尽管用户信任与求助行为呈正相关，但AI所带来的感知威胁会负向调节这一关联。本研究不仅丰富了关于生成式AI特征及其对心理健康求助行为影响的理论探讨，同时也为优化AI技术的设计与应用以提升心理健康支持效果提供了实践参考。