Trust in artificial intelligence, minority stress, and suicidal behavior severity among Black sexual minority male adolescents

Abstract Background Suicide is a leading cause of death among U.S. adolescents, with increasing concern about suicide risk among Black youth and sexual minority youth. Black sexual minority male (BSMM) adolescents remain markedly understudied, particularly at the intersection of racism, sexuality-related marginalization, and emerging sociotechnical environments. Guided by minority stress theory and a sociotechnical lens, this study examined associations among discrimination-related stressors, trust in artificial intelligence (AI), and suicidal behavior severity among BSMM adolescents ages 14–17 years. Methods Cross-sectional web-based survey data were analyzed from BSMM adolescents recruited from three Midwestern U.S. cities. The full available sample included 384 adolescents, with an analytic sample of 371 in the fully adjusted regression models. Suicidal behavior severity was assessed using a study-created composite adapted from five self-report suicidal behavior items from the Columbia–Suicide Severity Rating Scale. Discrimination-related stressors included day-to-day enacted discrimination, anticipated discrimination, and racial discrimination. Trust in AI was measured using a 7-item scale assessing perceived reliability, fairness, accuracy, and trustworthiness. Hierarchical linear regression models with robust standard errors were estimated in sequential blocks: sociodemographic characteristics, discrimination-related stressors, and trust in AI. Results In bivariate correlations, suicidal behavior severity was positively associated with day-to-day discrimination, anticipated discrimination, and racial discrimination, and negatively associated with trust in AI. In hierarchical regression models, the addition of discrimination-related stressors significantly improved model fit, ΔR² = 0.218, F(3, 360) = 50.05, p < .001. Day-to-day enacted discrimination (B = 0.59, p = .001) and racial discrimination (B = 0.66, p < .001) were independently associated with higher suicidal behavior severity in the fully adjusted model, while anticipated discrimination was not significant. Food insecurity was also associated with higher suicidal behavior severity (B = 1.20, p < .001). Trust in AI was negatively associated with suicidal behavior severity in the fully adjusted model (B = -0.22, p = .033). Conclusions Discrimination-related stressors, food insecurity, and trust in AI were associated with suicidal behavior severity among BSMM adolescents ages 14–17 years. Findings underscore the importance of addressing racism, day-to-day discrimination, and unmet basic needs in suicide prevention efforts with BSMM adolescents, while also highlighting trust in AI as a sociotechnical factor that warrants further study. Future research should examine AI trust and AI use simultaneously to clarify how emerging technologies shape mental health equity and suicide-related outcomes among marginalized youth.