Insomnia as an Independent Behavioral Correlate of Estimated Type 2 Diabetes Risk: Evidence from Three Validated Risk Assessment Scales.

In recent years, generative artificial intelligence (GenAI) has rapidly permeated all areas of higher education, to such an extent that it is now an indispensable part of university students’ academic lives [,]. GenAI tools such as ChatGPT and DeepSeek are now widely used by university students for tasks ranging from drafting papers and solving complex problems to the generation of creative content [,]. Although GenAI offers unprecedented opportunities for improving learning efficiency and accessibility, its impact on students’ learning is known to be far from unidimensional [,]. On the one hand, GenAI can serve as a powerful cognitive scaffold that empowers personalized learning and academic success [,]; on the other hand, it can undermine critical thinking, lower cognitive engagement, and foster passive learning behaviors [,]. GenAI thus has complex and multifaceted associations with learning, pointing to a need for a detailed and in-depth examination of its educational implications. The classroom is the primary setting where structured learning takes place, and the teaching interactions that occur within this space is significantly associated with students’ academic development. In traditional university classrooms, insufficient student engagement has long been a concern [–]. In the absence of immediate feedback and personalized encouragement, students’ interest in learning and their self-confidence are likely to be adversely affected [,]. By contrast, GenAI technology offers core capabilities such as natural language interaction, real-time feedback, and personalized adaptation, and it thus provides a new pathway for tackling the challenges of traditional teaching. Classroom use of AI is becoming increasingly prevalent, and its impact is becoming more pronounced [,]. Whether introduced by instructors as a teaching aid or independently adopted by students, AI technology is currently reshaping the ecological structure of classroom learning []. A new triadic, interactive model encompassing students, teachers, and AI is gradually replacing the traditional dyadic, teacher-student interactive model as student-AI interaction becomes a primary form of classroom participation []. Against this backdrop, can the integration of AI into classroom learning promote students’ learning engagement? Studies on the relationship between AI usage and learning engagement have mostly been conducted in experimental or quasi-experimental settings with specific teaching interventions [], or have not been confined to in-class learning contexts (e.g., []). With the ongoing popularization of AI usage, university students are increasingly turning to AI tools on their own initiative, even in the absence of conscious guidance from teachers, to assist their learning in classrooms [,]. The current study directly focuses on in-class AI usage and in-class learning engagement. It explores the status quo of university students’ attitudes toward in-class AI usage and their perceptions of whether in-class AI usage can enhance learning engagement, whether learning engagement mediates the relationship between AI usage behavior and attitudes, and whether there are differences in those mediating effects across the various dimensions of learning engagement. This study aims to provide empirical evidence and decision-making reference for the rational understanding and effective guidance of AI technology use to empower classroom teaching.