<![CDATA[The rapid adoption of generative artificial intelligence (generative AI) in higher education has raised concerns about its impact on students' critical thinking in academic writing, yet evidence on the conditions under which the two are related remains inconsistent. This quantitative correlational study examined the relationship between generative AI use and critical thinking among 34 Indonesian Language and Literature Education (PBSI) students participating in the Merdeka Student Exchange Program (PMM) when writing scientific articles. Drawing on Social Cognitive Theory and the cognitive offloading perspective, generative AI use was measured through four indicators (frequency, purpose, dependence, and AI literacy), and critical thinking through five aspects (analysis, evaluation, inference, synthesis–explanation, and reflection), using a four-point Likert questionnaire. Data were analyzed using simple and multiple linear regression, an independent-samples t-test, and moderated hierarchical regression. Composite generative AI use was not linearly related to critical thinking (p = .131) but showed an inverted-U pattern (p = .010), and intensive and selective users did not differ significantly (p = .097). AI literacy moderated the relationship between usage intensity and critical thinking (?R² = .222, p = .004; Holm-adjusted p = .016), with a significant positive relationship only among students with above-average AI literacy. These findings indicate that strengthening AI literacy in academic writing instruction, rather than merely restricting generative AI use, should underpin responsible, literacy-based generative AI policies in higher education.]]>
