Unethical Generative AI Use and Academic Procrastination in Higher Education: A Cross-Sample Examination of Epistemic Laziness and Cognitive Offloading

The increasing use of generative artificial intelligence (GenAI) in higher education has raised concerns about maladaptive academic behaviors. Using cross-sectional data from two independent university student samples, this study examined whether unethical GenAI use (UGAIU) was related to academic procrastination directly and through sequential indirect associations via epistemic laziness and cognitive offloading. Turkish-speaking students were surveyed in classrooms (N = 243) and English-speaking students online via Prolific (N = 213). Because the samples differed in several respects, they were treated as independent samples rather than as proxies for language or culture. The newly developed measures showed configural and metric invariance (only tentatively for epistemic laziness). Structural equation models controlling for gender, age, and daily GenAI use showed that UGAIU was positively associated with epistemic laziness and cognitive offloading in both samples. In the Turkish-speaking sample, UGAIU was indirectly associated with procrastination, mainly through cognitive offloading; in the English-speaking sample, only the direct association was significant. Path-specific multigroup tests provided weak evidence of a between-sample difference only for the cognitive offloading–procrastination path, which was not robust to multiple-comparison correction or partial invariance. Given the cross-sectional design, the findings describe associations rather than temporal or causal ordering.