Cognitive and attitudinal correlates of trust in AI-generated content among university students in Türkiye
We examined cognitive and attitudinal correlates of trust in AI-generated content in a 2025 cross-sectional survey in Türkiye. The respondent cohort comprised 232 records; 207 self-identified generative AI users completed the focal measures. A hierarchical regression distinguished perceived Generative AI Literacy, behavioral use frequency, positive and negative AI attitudes, an exploratory post-truth index, and risk/regulatory concern. Literacy and frequency explained 0.2% of trust variance; attitudes added 12.2%, and the joint post-truth/concern block added 13.6%. The final model explained 26.0% (adjusted R² = 0.238). Positive attitude (β = 0.310), negative attitude (β = −0.215), and the post-truth index (β = 0.425) had significant conditional associations. Negative attitude had an essentially null zero-order association ( r = .021). The post-truth index uniquely accounted for 13.4% points beyond the other measured predictors. Exploratory structure analyses supported a dominant common dimension but also revealed residual item clustering; content-domain sensitivity analyses localized the association mainly to devaluation of factual truth. Education-group differences did not survive adjustment across six outcome tests. These self-report associations do not establish causality, objective misinformation susceptibility, or a validated unitary post-truth disposition.
Authors
- Murat Aytaş (ORCID: https://orcid.org/0000-0003-2744-0519)
- Mehmet Kucuktongur
Institutions
- Selçuk University (TR)
- Ağrı İbrahim Çeçen University (TR)
Publication Details
- Journal
- Scientific Reports
- Published
- 2026-11-19
- DOI
- https://doi.org/10.1038/s41598-026-78416-2
- Primary Topic
- Misinformation and Its Impacts
- Type
- article
- Field-Weighted Citation Impact
- 0.00