This scoping review examines the influence of artificial intelligence (AI), particularly large language models (LLMs), on the effective use of critical thinking (CT) in higher education. Using Preferred Reporting Items for Systematic reviews and Meta‑Analyses extension for Scoping Reviews (PRISMA ScR) guidelines, a systematic search of Web of Science in March 2025 for English language, peer reviewed studies published between 2021 and early 2025 identified 44 articles for thematic synthesis. The review highlights three key themes: conceptualizations of CT, assessment methodologies, and potential risks. Findings indicate that AI tools can enhance skills such as inference, bias detection, creativity, self efficacy, and problem solving when combined with domain specific scaffolding and metacognitive strategies, yet they also pose risks including surface level learning, stagnation in cognitive growth, and challenges to academic integrity. The necessity for deliberate instructional design and ethical awareness when incorporating AI into curricula is underscored. Recommendations include aligning pedagogy with a seven step CT model and employing purpose and context fit, hybrid assessment strategies. The review’s focus on English language, peer reviewed journal articles and the exclusion of grey literature, combined with the absence of a formal quality appraisal, may constrain the comprehensiveness and generalisability of its findings. Ultimately, LLMs are positioned as catalysts for deeper cognitive engagement, fostering reflective and adaptive reasoning in complex, algorithmically influenced educational landscapes.

Keywords: Artificial Intelligence in Education, Critical Thinking Skills, Higher Education.