BEYOND CHATGPT: EXAMINING THE IMPACT OF GENERATIVE ARTIFICIAL INTELLIGENCE LITERACY ON HIGHER ORDER THINKING SKILLS THROUGH SELF-REGULATED LEARNING AMONG UNIVERSITY STUDENTS

Abstract

University classrooms did not so much adopt generative artificial intelligence as get overtaken by it, and the pedagogical thinking has been trying to catch up ever since. This narrative review looks at how the different dimensions of generative AI (GenAI) literacy relate to university students' Higher Order Thinking Skills (HOTS), and more specifically at the part Self-Regulated Learning (SRL) plays in deciding whether a chatbot ends up sharpening a student's thinking or quietly doing that thinking for them. Literature was drawn mainly from Scopus- and Web of Science-indexed journals, several arXiv preprints, and major international survey reports published between roughly 2023 and 2026, then organised around four anchors: AI literacy frameworks, Bloom's Revised Taxonomy, Zimmerman's cyclical SRL model, and the empirical record on cognitive offloading and academic integrity. Read together, this literature treats GenAI literacy as a multidimensional construct covering technical, critical-evaluative, ethical, and metacognitive competencies, and it treats the GenAI-HOTS relationship as conditional rather than direct: the same tool produces very different outcomes depending on whether its use is structured and metacognitively guided or left unstructured. SRL is what tips the balance. It is the variable that turns GenAI into either a scaffold for analysis, evaluation, and creation, or a substitute that lets those skills quietly atrophy. The review closes with implications for curriculum design, arguing that GenAI literacy and SRL need to be taught together, deliberately, rather than assumed to develop on their own from mere exposure, and it sketches a research agenda addressing the field's continuing shortage of longitudinal and non-Western evidence.