The Social Anatomy of AI Anxiety: Gender, Generations, and Technological Exposure

Naciye Güliz UğurFaruk Dursun^*

• Sakarya University, Sakarya, Türkiye

Public anxiety about artificial intelligence (AI) has policy, clinical, and educational implications, yet evidence on its multidimensional structure and population correlates remains fragmented. This study synthesizes validated measures into a nine-dimension framework to characterize AI-related anxiety and its demographic and experiential determinants. A cross-sectional survey of N = 1,151 adults captured nine anxiety dimensions—general AI anxiety, technoparanoia, technophobia, AI interaction anxiety, job-replacement anxiety, sociotechnical blindness, cybernetic-revolt fear, technology self-efficacy, and AI learning orientation—adapted from established scales (not a new instrument). Dimensionality was examined via common-factor EFA (principal axis factoring, Promax; KMO = .89; Bartlett p < .001) with parallel analysis and scree inspection. A hold-out CFA (70/30 split) assessed model fit. Reliability (α, ω), composite reliability (CR) and average variance extracted (AVE) tested internal consistency and convergent validity; discriminant validity used Fornell–Larcker and HTMT. Group differences (t-tests/ANOVA) reported effect sizes (Cohen's d, partial η²) with Holm–Bonferroni control of family-wise error. Predictors of overall AI anxiety were evaluated with hierarchical regression controlling for age, gender, marital status, employment, and AI-use status. EFA supported a nine-factor solution (64.17% variance). CFA indicated acceptable fit (CFI = .943, TLI = .936, RMSEA = .045 [90% CI .041–.049], SRMR = .046). All factors showed α, ω ≥ .80, CR ≥ .83, AVE ≥ .51, and satisfactory discriminant validity. After correction, gender differences persisted for technoparanoia, AI learning orientation, and interaction anxiety (small effects; d ≈ .18–.21). AI users scored higher on general AI anxiety, technoparanoia, and sociotechnical blindness (d ≈ .17–.29), while age-group differences were non-significant across dimensions. The hierarchical model controlling for age, gender, marital status, employment, and AI usage explained R² = .512 of AIA: sociotechnical blindness and technoparanoia were the strongest positive predictors; technology self-efficacy and AI learning orientation were negative predictors; job-replacement anxiety contributed positively with a smaller effect. Demographic effects were modest after psychological covariates entered. AI-related anxiety is psychometrically robust, multidimensional, and primarily psychologically driven rather than demographic in origin. Findings highlight actionable levers for intervention—boosting self-efficacy, structured AI literacy, and transparent sociotechnical risk communication—to mitigate anxiety while enabling informed, equitable AI integration in clinical, educational, and policy settings.