Study: Undergraduates Conceal Generative AI Use From Instructors; Fear of Teacher Judgment Correlates With Peer Stigma

Researchers from Simon Fraser University and partner institutions published an original study in Frontiers in Education based on survey data from 78 undergraduates in an online education course, examining their generative AI disclosure practices. The study found that fear of teacher judgment and fear of social stigmatization were strongly correlated, and both were associated with students actively concealing AI use; anxiety was not directly linked to reduced disclosure to teachers but was associated with more peer-only sharing and complete secrecy. The authors treat the results as hypotheses for future work given the limited sample size.

The study offers student-behavior evidence for universities drafting AI disclosure rules: if fear of teacher judgment and peer stigma coexist, compliance requirements alone may not produce honest AI-use declarations and may instead push students toward peer-only sharing or complete secrecy. For teachers and schools, this means disclosure systems need to account for trust and evaluation pressure; for AI education product teams, how in-product usage records and declaration features connect to course assessment directly affects whether students leave truthful data. The study covers only 78 students in a single online course, and the authors themselves frame the conclusions as hypotheses to be tested, so it should not be used to infer general learning outcomes or institutional effectiveness. Universities that rely only on compliance-based AI disclosure may not raise disclosure rates, as students still turn to peer-only sharing or complete secrecy. Teachers and course designers need to factor evaluation pressure and peer stigma into disclosure mechanisms, or they will struggle to obtain accurate AI-use information.