The increasing use of generative artificial intelligence (GenAI) has transformed students’ approaches to academic writing. This study explores how university students integrate GenAI into academic writing and how sustained use relates to reliance, trust, verification, and self-regulation. A qualitative exploratory case study was conducted through semi-structured interviews with three Indonesian undergraduate EFL students (P01, P02, and P03), with the data analyzed using reflexive thematic analysis. Four interconnected themes emerged. First, GenAI had developed from a language-support tool into an integrated resource across planning, drafting, and revision. Second, its use shifted from occasional assistance to more routine reliance, particularly when students encountered linguistic or organizational difficulties. Third, participants demonstrated selective trust, accepting linguistic assistance more readily while verifying factual claims, theoretical explanations, and references against academic sources. Fourth, GenAI facilitated monitoring and revision but was also associated with reduced independent exploration and tolerance for prolonged difficulty. Despite this emerging reliance, participants retained control over major academic decisions and remained capable of writing without AI. The findings suggest that GenAI use is better characterized as functional reliance rather than complete cognitive dependence, with self-regulation, verification, and critical evaluation serving as important conditions for maintaining students’ cognitive responsibility in academic writing
Copyrights © 2025