Methodology for Using ChatGPT to Develop Critical Thinking Skills in Higher Education Students

Abstract

The purpose of the article is to develop and provide a theoretical substantiation for a methodology of using ChatGPT in the educational process of higher education institutions to foster students’ critical thinking at the levels of analysis, evaluation, and creation within Bloom’s taxonomy. The relevance of the study is determined by the contradiction between the rapid integration of generative artificial intelligence into higher education and the insufficient methodological preparedness of educators to use this tool in ways that do not diminish cognitive engagement but, on the contrary, stimulate students’ intellectual independence. The research methods include the analysis and synthesis of recent scholarly studies on the use of ChatGPT in higher education; comparison of approaches to developing higher-order thinking skills; generalisation of the findings of systematic reviews and empirical studies; modelling of instructional techniques for integrating ChatGPT into classroom activities; and designing an algorithm of students’ critical interaction with a generative model. The results show that ChatGPT is most effective as a tool for supporting lesson planning, developing learning materials, providing timely feedback, and personalising instruction; however, in the absence of deliberate instructional design, it may encourage superficial consumption of ready-made answers, weaken independent argumentation, and promote the uncritical acceptance of inaccurate statements and fabricated sources. On this basis, the article proposes a methodology that treats ChatGPT not as a source of final knowledge but as an instrument of Socratic dialogue, an intellectual interlocutor, and an object of verification. The methodology includes a typology of tasks structured according to Bloom’s taxonomy; an algorithm of critical interaction with artificial intelligence; three core instructional techniques—“hallucination verification,” “comparative prompt analysis,” and “AI as reviewer”; and principles for the ethical, transparent, and verifiable use of AI-generated content. It is demonstrated that the pedagogical value of ChatGPT increases under conditions of human oversight, clear prompt design, reliance on scholarly sources, and specially organised tasks in which students verify, compare, evaluate, and refine generated responses. The conclusions suggest that the proposed methodology is aimed at minimising the risk of cognitive passivity and excessive dependence on generative artificial intelligence and can be applied in the teaching of humanities, social sciences, economics, and natural sciences in blended, face-to-face, and mobile-supported learning environments.