Study: 7 psychological factors shape students' honest GenAI use

A new academic paper from researchers Ezgi Dagtekin and Ercan Erkalkan—presented at the 11th International Academic Studies Congress in Mersin, Türkiye, and posted to arXiv (arXiv:2608.14605)—reframes GenAI-related academic misconduct as a psychologically mediated decision process. Rather than blaming technology, the authors synthesize evidence from 16 peer-reviewed and policy publications (2022–March 2026) to identify six psychological determinants: moral reasoning, perceived social norms, policy clarity, academic self-efficacy, AI literacy, performance pressure, and beliefs about authorship. The review finds that students do not uniformly equate AI assistance with cheating; integrity risk spikes when institutional guidance is vague, peers appear to use AI without consequence, and AI tools are seen as legitimate substitutes for difficult cognitive labor.

The paper’s integrative conceptual model connects institutional context to psychological appraisal, which then leads to disclosed, borderline, or dishonest GenAI use. Encouragingly, the evidence shows that assignment-level guidance, explicit disclosure norms, ethics-focused instruction, and authentic assessment design reduce integrity risk more effectively than detection-centric surveillance. The authors conclude that universities should respond with a blend of clear policy, pedagogy, AI literacy, and student support—not prohibition or software monitoring alone. The 10-page paper includes a figure and was also published in the congress Book of Full Texts (pp. 750–759). This gives educators an evidence-based roadmap to foster honest AI adoption without relying on punitive measures.