Artificial Intelligence (AI) is reshaping higher education by enabling personalized learning (PL) and enhancing teaching and learning practices. To examine global research trends, pedagogical paradigms, equity and sustainability considerations, instructional strategies, learning outcomes, and interdisciplinary collaboration, this study systematically reviewed 29 articles indexed in the Social Sciences Citation Index (SSCI) Q1, representing the top 25% of cited articles, published between January 2020 and December 2024 in the Web of Science database. Results indicate that AI-PL research is concentrated in Asia, particularly China, and predominantly situated within education and computer science. Quantitative designs prevail, often complemented by qualitative insights, with supervised machine learning as the most common algorithm. While constructivist principles implicitly guide most studies, explicit theoretical grounding improves AI-pedagogy alignment and educational outcomes. AI demonstrates potential to enhance instructional approaches such as PBL, STEAM, gamification, and UDL, and to foster higher-order skills, yet uncritical use may undermine learner autonomy. Systematic attention to equity and SDG-related objectives remains limited. Emerging interdisciplinary collaborations show promise but are not yet fully institutionalized, constraining integrative system design. These findings underscore the need for stronger theoretical framing, alignment of AI with pedagogical and societal imperatives, and professional development to enhance educators’ AI literacy. Coordinated efforts among academia, industry, and policymakers are essential to develop scalable, context-sensitive AI solutions that advance inclusive, adaptive, and transformative higher education.

Project-based learning, in which students engage in meaningful learning with authentic projects and building agency and autonomy for sustainable learning, has been increasingly promoted in higher education. However, completing an authentic project involves a complex process, which may pose challenges to many students, especially low-achievers. This study incorporated computer-based scaffolding into a project-based programming course to make complex project learning accessible to students. The scaffolding was designed based on the four-component instructional design (4C/ID) model. The results show that with the support of computer-based scaffolding, all participants maintained a high level of motivation during the course. At the end of the course, their performance was improved by 35.49% in product quality and 38.98% in subject knowledge; their programming thinking skills were improved by 20.91% in problem understanding, 21.86% in modular design, and 25.70% in process design. Despite academic achievement discrepancies among the participants at the beginning of the course, low-achievers’ post-study performance in product quality and programming thinking skills became similar to that of high-achievers, and their post-study performance in subject knowledge became similar to that of medium-achievers. The findings reveal the promising role of computer-based scaffolding in making complex learning with real-world projects accomplishable by a wide range of students and reducing the gaps between high- and low-achieving students.

Background: The provision of public adaptive coping strategies to reduce psychological tension during the ongoing COVID-19 pandemic is critical. We sought to provide evidence-based guidance for psychological intervention, exploring the potential mediating roles of three sources of social support (i.e., subjective support, family support and counselor support) between coping strategies (i.e., cognitive coping, emotional coping and behavioral coping), and anxiety among college students at the height of the pandemic in China. Methods: Using the Coping Strategy Questionnaire, Social Support Questionnaire, and Self-Rating Anxiety Scale, this large-scale online study analyzed the levels of social support, coping, and anxiety among 2640 college students in China from February 21st to 24th, 2020, when the students had been isolated at home for 1 month since the lockdown of Wuhan city. Results: Students reported high levels of cognitive coping, behavioral coping, and social support. They also experienced low levels of anxiety and emotional coping. Anxiety was significantly and negatively related to coping and social support. The mediating roles of three sources of social support were found between cognitive coping, behavioral coping, and anxiety, respectively. However, the effect of emotional coping on anxiety was not found to be mediated by social support. Conclusions: Adopting positive coping strategies may enhance social support that in turn relieves anxiety. The effect of social support, especially family and counselor support, should arouse greater awareness in coping with the pandemic cognitively and behaviorally.

With information technology pushing the development of intelligent teaching environments, the online teaching platform emerges timely around the globe, and how to accurately evaluate the effect of the “any-time and anywhere” teacher–student interaction and learning has become one of the hotspots of today’s education research. Bullet chatting in online courses is one of the most important ways of interaction between teachers and students. The feedback from the students can help teachers improve their teaching methods, adjust teaching content, and schedule in time so as to improve the quality of their teaching. How to automatically identify the sentiment polarity in the comment text through deep machine learning has also become a key issue to be automatically processed in online course teaching. The traditional single-layer attention mechanism only enhances certain sentimentally intense words, so we proposed a sentiment analysis method based on a hierarchical attention mechanism that we called HAN. Firstly, we use CNN and LSTM to extract local and global information, gate mechanisms are used for extracting sentiment words, and the hierarchical attention mechanism is then used to weigh the different sentiment features, with the original information added to the attention mechanism concentration to prevent the loss of information. Experiments are conducted on China Universities MOOC and Tencent Classroom comment data sets; both accuracy and F1 are improved compared to the baseline, and the validity of the model is verified.