This study investigates the impact of AI-assisted language learning (AIAL) strategies on cognitive load and learning outcomes in the context of language acquisition. Specifically, the study explores three distinct AIAL strategies: personalized feedback and adaptive learning, interactive exercises with speech recognition, and intelligent tutoring with data-driven insights. The research employs a pretest-posttest random assignment experimental design, utilizing three experimental groups and a control group, with a total of 484 EFL students specializing in teaching English as a foreign language participating in the study. Data collection involves pre- and post-tests, questionnaires, and interviews to assess the influence of AIAL strategies on cognitive load and learning outcomes. Cognitive load is measured using the Cognitive Load Scale, while pretest-posttest assessments evaluate the efficacy of AIAL interventions across various language skills. These results contribute to the existing body of AIAL research by offering empirical evidence for the effectiveness of specific strategies in optimizing language learning experiences. The implications of this study extend to educators, researchers, and developers in the field of AIAL, emphasizing the potential of AIAL to enhance language acquisition processes and inform instructional design practices.

This study presents an inclusive K-12 AI curriculum for elementary schools, focusing on six design principles to address gender disparities. The curriculum, designed by the researchers and an elementary teacher, uses tangible tools, and emphasizes collaboration in solving daily problems. The MANOVA results revealed initial gender differences in AI knowledge. Following MANCOVA analysis showed significant improvements in AI knowledge, as well as perceptions and behavioral intentions toward AI, effectively bridging the observed knowledge gaps without any significant differences attributed to gender or initial knowledge levels. This evidence underscores the success of tangible and collaborative AI interventions in uniformly enhancing AI knowledge and positively changing perceptions and behavioral intentions among all participants, regardless of gender. Both female and non-binary students felt increased engagement and reduced anxiety toward AI, with improved comprehension and a preference for collaborative learning. This study contributes to the design of inclusive AI interventions, highlighting gender differences and including non-binary students as an essential part of the analysis.

Scholarly interest in artificial intelligence (AI) has surged as researchers delve into its transformative impact on various aspects of our lives. AI poses both benefits and challenges, particularly in the context of educators' endeavors to comprehend the intricacies of students' learning processes. Although the use of AI to enhance and assist student learning is relatively new, the exponential growth of scholarly attention and publications in AI and student learning in recent years underscores the compelling necessity for further inquiry. Investigating this area is crucial for understanding the emerging trends in this research domain. This study aims to provide insights into the burgeoning research trajectories on AI from a student learning perspective. Using a bibliometric approach, this study examined 663 scholarly articles pertaining to the interface between AI and student learning published between 1961 and 2024. Our findings reveal four major thematic areas including AI in education and educational technology, AI-driven learning environments, essential AI enablers, and human learning and highlight promising avenues at this intersection.

The use of generative artificial intelligence (Gen-AI) to assist college students in their studies has become a trend. However, there is no academic consensus on whether Gen-AI can enhance the academic achievement of college students. Using a meta-analytic approach, this study aims to investigate the effectiveness of Gen-AI in improving the academic achievement of college students and to explore the effects of different moderating variables. A total of 28 articles (65 independent studies, 1909 participants) met the inclusion criteria for this study. The results showed that Gen-AI significantly improved college students’ academic achievement with a medium effect size (Hedges’s g = 0.533, 95% CI [0.408,0.659], p < .05). There were within-group differences in the three moderator variables, activity categories, sample size, and generated content, when the generated content was text ( g = 0.554, p < .05), and sample size of 21–40 ( g = 0.776, p < .05), the use of independent learning styles ( g = 0.600, p < .05) had the most significant improvement in college student’s academic achievement. The intervention duration, the discipline types, and the assessment tools also had a moderate positive impact on college students’ academic achievement, but there were no significant within-group differences in any of the moderating variables. This study provides a theoretical basis and empirical evidence for the scientific application of Gen-AI and the development of educational technology policy.

Today, immersive technologies like Virtual Reality (VR) are regarded as disruptive tools in many domains, including education. While the body of literature in the field is growing, studies that present contrasting findings are not uncommon. In fact, although there is evidence of the benefits brought by VR in the educational processes, in some cases the effects of a possible trade-off between learning effectiveness and quality of the learning experience (or QoLE) may be observed. The two dimensions are difficult to disentangle, as besides learning effectiveness, other factors like motivation, technology acceptability, workload, presence, immersion, engagement, and usability come to play. This paper digs into the above scenario by focusing on the QoLE of immersive VR-based learning and comparing it with that of two conventional approaches (a physical prop-based one and a 3D desktop application). Separation of the two dimensions is pursued by imposing equality of the learning performance achieved with the three approaches, aiming at getting rid of possible confounding factors. From the results of the user study performed in the context of a STEM-related laboratory activity, the VR-based approach appeared to be generally superior to the prop-based approach and showed several advantages over the 3D desktop application.

The study investigates the potential of anxiety clusters in predicting programming performance in two distinct coding environments. Participants comprised 83 second-year programming students who were randomly assigned to either a block-based or a text-based group. Anxiety-induced behaviors were assessed using physiological measures (Apple Watch and Electrocardiogram machine), behavioral observation, and self-report. Utilizing the Hidden Markov Model and Optimal Matching algorithm, we found three representative clusters in each group. In the block-based group, clusters were designated as follows: “stay calm” (students allocating more of their time to a calm state), “stay hesitant” (students allocating more of their time to a hesitant state), and “to-calm” (those allocating minimal time to a hesitant and anxious state but displaying a pronounced propensity to transition to a calm state). In contrast, clusters in the text-based group were labeled as: “to-hesitant” (exhibiting a higher propensity to transition to a hesitant state), “stay hesitant” (allocating significant time to a hesitant state), and “stay anxious” (remaining persistently anxious in a majority of the coding time). Additionally, our results indicate that novice programmers are more likely to experience anxiety during text-based coding. We discussed the findings and highlighted the policy implications of the study.

The present research conducted two experiments with an intelligent tutoring system to investigate the overall and dynamic impact of emotional support from a pedagogical agent (PA). In Experiment 1, a single factor intergroup design was used to explore the impact of PA’s emotional support (supportive vs. non-supportive) on learners’ emotions, intrinsic motivation, and learning gain. Sixty participants were recruited and randomly assigned to one of the two conditions. Experiment 2 also conducted a single factor between-subjects design to investigated the dynamic patterns between emotions and learning strategies among 30 participants using lag sequential analysis. Results showed that: Compared with the non-supportive pedagogical agent, the supportive pedagogical agent reduced frustration and improved learning gain, but did not increase intrinsic motivation. In addition, learners with the supportive pedagogical agent used more appropriate strategies after frustration and surprise, and use less ineffective strategies after confusion and enjoyment to avoid reaching a wrong answer. If learners did not receive emotional support in such cases, learning strategies following these emotions were more likely lead to negative cognitive results, or negative emotions tended to appear repeatedly. Instructors or PAs should identify learners’ emotions in time and provide the appropriate emotional support according to learners’ emotions.