Abstract Despite artificial intelligence (AI)’s transformative impact on education, empirical work on students’ ethical awareness and actual practices when navigating AI-related dilemmas remains scarce. This study examined 25 Chinese doctoral students across disciplines via an AI-assisted simulated literature review task, two qualitative surveys, and student-AI dialogues. Key findings: (1) Most demonstrated AI ethics awareness but focused on system-related concerns over self-relevant issues (e.g. eroded autonomy, blurred authorship). (2) All students used AI in at least one writing phase (literature understanding, synthesis, drafting, and editing) for one or more purposes (conceptual, rhetorical, or linguistic support). Faced with ‘disclose or not’ dilemma, most opted for partial disclosure in editable text. Individual variation was however observed in distributed agency and disclosure extent. (3) Discrepancies were identified between students’ stated ethical beliefs and actual behavior when navigating ethical dilemmas. Pedagogical implications were discussed including an advocate for creating safe havens for students to experiment AI-use and negotiate context-specific ethics.
Feng et al. (Thu,) studied this question.