AI-based early mental health screening for local public officials

Local public officials are consistently exposed to high levels of occupational stress and burnout due to heavy workloads, citizen interactions, and rigid organizational cultures. Conventional self-report assessments are limited by subjectivity and stigma, whereas artificial intelligence (AI)-based screening systems offer potential advantages for earlier engagement and improved accessibility, particularly in stigma-sensitive contexts. This study evaluated the acceptability and perceived usefulness of an AI-based mental health screening system for local officials.

A two-stage survey was conducted with 126 officials in local public. Stage one employed standardized instruments (PHQ-9, GAD-7, job stress, stigma perception). Stage two included 30 randomly selected participants who completed both self-report and GPT-based AI assessments, followed by evaluation of satisfaction and acceptance. Analyses included descriptive statistics, ANOVA, Chi-square tests, reliability (Cronbach's α), correlation, and multiple regression using SPSS 29.0.

Respondents reported high job stress, with younger officials showing greater vulnerability. Despite only 3.2% having prior AI diagnostic experience, 73.8% trusted AI results more than self-reports, and 84.9% expressed willingness for regular participation. Stigma toward conventional services remained prevalent, yet 81.7% agreed AI could mitigate stigma. In the in-depth survey, AI screening was rated highly for usability, efficiency, interpretability, and re-engagement intention, although some concerns regarding transparency and interpretability remained. Therefore, the AI model in this study should not be interpreted as providing clinical diagnoses. Its outputs reflect pattern recognition within conversational data, and further validation against clinician-administered psychiatric assessments is required. Regression analysis showed that trust in AI results (β = 0.463, p < 0.001) and perceived stigma reduction (β = 0.432, p < 0.001) significantly predicted AI acceptance.

AI-based screening demonstrates strong potential as a perceived potential and stigma-free tool for public officials. To ensure sustainable implementation, institutional safeguards and algorithm refinement are required.