This paper investigates how higher education students reflected on their use of ChatGPT to support critical reading and source evaluation within a higher education digital literacy course in Greece. Drawing on students’ written reflections, forum discussions, collaborative tasks and questionnaires, the study explores how students interacted with ChatGPT as a dialogic partner for questioning, verification, and reasoning. Thematic analysis of the students’ data identified four interrelated themes: (1) prompting as a metacognitive practice, (2) developing a critical stance through evaluative judgment of ChatGPT’s outputs, (3) recognizing the tool’s conditional usefulness, and (4) addressing language-related challenges. Findings show that students demonstrated awareness of how the formulation of prompts influenced response quality and depth, using inaccuracies and generalizations as opportunities for deeper analysis. They valued ChatGPT’s immediacy and accessibility but maintained a cautious stance toward its reliability, highlighting the need for verification and human mediation. Differences in performance across languages also prompted reflection on linguistic precision and bias. Overall, students’ reflections highlight that structured, reflective prompting can foster metacognitive regulation, evaluative judgment, and critical reading, positioning ChatGPT as a catalyst for inquiry and ethical awareness rather than a substitute for human reasoning.
Key words: ChatGPT; critical reading; source evaluation; metacognition; AI literacy; higher education; language awareness