Cognitive Load in Generative AI-Supported Learning in Higher Education: A Systematic Review

Generative AI is increasingly used in higher education, but how GenAI-supported interaction relates to students’ cognitive load-related experiences remains unclear. This review examined empirical studies published between 2022 and 8 May 2026 that investigated higher education students’ interaction with GenAI systems and cognitive load-related experiences. Twenty-one studies met the inclusion criteria, covering diverse learning contexts including language and writing, STEM, professional education, design, and self-regulated learning. Across studies, findings revealed diverse cognitive patterns rather than a consistent cognitive outcome of GenAI-supported learning. Some studies suggested that GenAI was associated with reduced unnecessary cognitive demands through explanation, feedback, and support. The findings suggest that reduced cognitive demand may reflect effective cognitive support, while some cases may indicate possible cognitive offloading when learners do not actively engage with AI-generated information. Overall, GenAI may provide greater learning value when interaction design reduces unnecessary difficulty while maintaining learners’ active engagement in evaluation and reflection.