This scoping review maps AI-based approaches used to infer or measure attention and emotion in technology-enhanced learning (TEL), with a particular focus on tertiary (higher) education and learning analytics-enabled digital environments supporting online and hybrid instruction. Although artificial intelligence (AI) promises personalized digital education, many systems still respond poorly to students’ attentional and emotional fluctuations. We therefore examined the extent to which the literature converges on jointly measuring attention and emotion through AI in educational contexts, especially in virtual and distance-learning settings. Following PRISMA-ScR, we searched Scopus and Web of Science and identified 39 eligible studies. We conducted a methodological quality appraisal using Joanna Briggs Institute tools, a keyword co-occurrence bibliometric analysis, and a narrative synthesis. The evidence shows a rapidly expanding field and a wide range of AI-based techniques, but emotion and attention are typically operationalized and modelled in isolation. Both the bibliometric and narrative results indicate persistent conceptual fragmentation and limitations in the validity of measurement metrics. Overall, the field has not yet established a unified paradigm that integrates attention and emotion within AI-driven educational systems, constraining their adaptive potential. This evidence highlights the need for theory-informed and operational frameworks that enable genuinely holistic, student-centred pedagogical adaptation.
Arranz-Romero et al. (Thu,) studied this question.