Beyond Frequency: Latent Profiles of Problematic ChatGPT Use and Psychosocial Risk and Protective Factors Among Adults
International Journal of Mental Health and Addiction, 2026 (SCI-Expanded, SSCI, Scopus)
- Yayın Türü: Makale / Tam Makale
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s11469-026-01712-5
- Dergi Adı: International Journal of Mental Health and Addiction
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Social Sciences Citation Index (SSCI), Scopus, Psycinfo, Social Sciences Abstracts, Academic Search Ultimate (EBSCO), Health Research Premium Collection (ProQuest)
- Anahtar Kelimeler: Artificial intelligence, ChatGPT, Latent profile analysis, Person-centered approach, Problematic ChatGPT use, Psychosocial risk
- Yıldız Teknik Üniversitesi Adresli: Evet
Özet
Problematic ChatGPT use (PCU) may depend less on frequency than on the psychosocial context in which engagement occurs. This study identified psychosocial profiles among 1051 adult ChatGPT users in Türkiye using latent profile analysis. Analyses were conducted using IBM SPSS Statistics 27 and R via RStudio, with profile estimation implemented using the mclust package. Seven indicators were included: problematic ChatGPT use, loneliness, intolerance of uncertainty, psychological flexibility, self-control, attitudes toward artificial intelligence, and perceived social support. Four profiles emerged: Lower PCU–Higher Resources, AI-Positive–Higher Resources, Lower PCU–Lower Resources, and Highest PCU–Lower Resources. The profiles differentiated higher-resource configurations varying in AI orientation from lower-resource configurations with contrasting PCU tendencies. Auxiliary analyses using an empirical BCH three-step approach to account for classification uncertainty showed profile differences in emotion-regulation strategies and all Big Five traits. Age, education, socioeconomic status, marital status, employment status, and ChatGPT use frequency differed across profiles, whereas gender did not. PCU was interpreted as self-reported dysregulation rather than a clinical diagnosis or independently verified functional impairment. Findings suggest that problematic ChatGPT use should be understood through broader psychosocial configurations rather than frequency alone. Future research should examine whether these profiles replicate across cultural contexts and population groups, particularly among adolescents and university students, and use longitudinal designs incorporating objective usage metrics and indicators of functional impairment to assess profile stability, transitions, and temporal ordering.