<![CDATA[Generative artificial intelligence (GenAI) is associated with diverse psychological experiences in university learning, but routine use and designed interventions require different interpretations. This critical review examines psychological outcomes, correlates, and direct resilience-related evidence in reports identified through searches of Scopus, ERIC, and the SpringerLink publisher platform on 18 September 2026. The recorded search yielded 1,013 records; 50 duplicates were removed, 963 records were screened, and 45 candidate reports were sought. A revision-stage retrieval audit increased assessed reports from 12 to 44. One report remained unavailable, and one was excluded because GenAI-specific exposure could not be separated from broader AI use. The synthesis therefore includes 43 reports: 37 observational or task-context reports, four educational intervention reports, and two psychological-support trials. Dependence, anxiety, efficacy beliefs, literacy, and regulation showed heterogeneous associations; neither frequent use nor anxiety uniformly indicated harm. A classroom comparison measured academic resilience, a survey measured academic buoyancy, and a single-group study examined buoyancy qualitatively. These findings correct the earlier inference that direct resilience-related evidence was absent. Intervention benefits were specific to programs, outcomes, and assessment periods. Non-probability sampling, self-report, confounding, heterogeneous measures, incomplete retrieval, and limitations in the original screening documentation preclude a comprehensive causal account. The authors propose counseling and teaching applications for evaluation rather than presenting established prevention or treatment effects.]]>
Copyrights © 2026
