The relationship between academic engagement and mental health has emerged as a critical area of research, particularly among college students, who often face significant academic and social pressures. Academic engagement refers to the level of interest, participation, and commitment that students demonstrate toward their studies, and it includes four key dimensions: affective, social, cognitive, and behavioral (Bowden, 2022; Bowden et al., 2021). These dimensions shape not only the academic experience but also student well-being. For Chinese college students, the importance of academic engagement is especially pronounced due to unique cultural and educational expectations. Research has consistently shown that high levels of academic engagement are associated with positive mental health outcomes, while disengagement can lead to stress, anxiety, and depression (Sinval et al., 2024; Tang and He, 2023).

Increased socioeconomic development and supportive national educational policies in China have played an important role in enhancing academic engagement, which in turn has positively influenced student mental health. For instance, Li F. et al. (2022) reported significant improvements in mental health among Chinese physical education college students from 1995 to 2019, attributing this improvement to rising levels of engagement fostered by socioeconomic growth and policy interventions. Despite the growing body of research on this topic, there remains a need for more empirical studies, particularly to explore the cross-cultural and context-specific impacts of academic engagement and stress on student mental health (Arterberry et al., 2024; Beroiza-Valenzuela, 2024; Dave et al., 2024).

In recent years, the integration of artificial intelligence (AI) in higher education has transformed both learning and teaching environments, particularly in the realm of mental health education (Li Z., 2023). This integration can be understood from two perspectives: the use of AI at the organizational level, where institutions and teachers implement AI-based interventions, and the use of AI by students in their own learning activities. From the organizational standpoint, AI has been instrumental in improving the effectiveness of mental health education through personalized, scalable, and accessible interventions, including cognitive-behavioral approaches that focus on enhancing students’ psychological well-being (Aminah et al., 2023; Wang et al., 2024). Wang’s et al. (2024) cognitive-behavioral model of psychological education, guided by AI, is one such example, which significantly increased student engagement and participation in mental health activities.

On the student side, the adoption of AI tools for learning activities has been explored through frameworks such as the Unified Theory of Acceptance and Use of Technology (UTAUT). This model emphasizes the importance of factors such as performance expectancy, effort expectancy, and social influence in shaping students’ use of AI technologies (Venkatesh et al., 2012). Recent research has suggested that AI adoption by students can improve their academic engagement, helping them cope with academic stress and enhance their mental health (Alqahtani et al., 2023; Zapata et al., 2024). However, criticisms of the UTAUT model have pointed out its limitations in addressing the social ramifications of new technologies and the complexities of AI adoption in educational contexts (Venkatesh et al., 2012). Additionally, the integration of AI in classrooms—such as in flipped classroom models—has been shown to boost engagement and improve student performance in mental health courses (Shan and Liu, 2021). While these benefits are clear, there are still concerns regarding the ethical implications of AI, including privacy and data security issues, as well as the risk of diminishing human interaction (Escotet, 2023).

These complexities can be understood through the theoretical framework of the Theory of Planned Behavior (TPB). According to TPB, behavior is influenced by three components: attitudes toward the behavior, subjective norms, and perceived behavioral control (Ajzen, 1991). In this context, positive attitudes toward AI in education, supportive social norms, and students’ confidence in using AI tools can significantly influence their intention to adopt AI. For example, attitudes toward AI have been shown to predict behavior in academic-related fields, including help-seeking (Bornschlegl et al., 2021), cheating intentions (Stone et al., 2010), and AI use among both teachers and students (Abdelmoneim et al., 2024; Duong et al., 2024). Studies have demonstrated that positive attitudes, favorable subjective norms, and high perceived control are associated with higher engagement with AI tools, which can improve both academic outcomes and mental health (Wang et al., 2024; You and Zhang, 2024).

Additionally, perceived behavioral control plays a significant role in determining whether students feel capable of using AI technologies effectively. Factors such as access to user-friendly applications, institutional support, and training programs can increase perceived behavioral control and, subsequently, AI adoption (Li X. et al., 2022; Shan and Liu, 2021). A supportive environment that addresses students’ attitudes, subjective norms, and perceived control can help foster positive behavioral intentions toward AI use, which may lead to improved academic engagement and mental health outcomes.

This study aims to explore how attitudes toward AI, subjective norms, perceived behavioral control, and behavioral intention influence the relationship between Chinese college students’ academic engagement and mental health. By integrating the Theory of Planned Behavior into this framework, the study seeks to answer the following research question: How do attitudes toward AI, subjective norms, perceived behavioral control, and AI use behavioral intention influence the relationship between Chinese college students’ academic engagement and their mental health?

By addressing this question, we aim to provide insights into how AI adoption can enhance the learning experience and support mental well-being among Chinese college students. The findings from this study will not only contribute to the academic understanding of AI’s role in higher education but also inform the development of effective policies and interventions. These could help leverage AI technologies to create more personalized, supportive, and engaging learning environments that promote holistic student development. Furthermore, the research has the potential to guide educational institutions in promoting the effective use of AI to foster active learning, increase academic engagement, and improve student mental health.

The following Figure 1 depicts the mediated chain model of relationships between the variables.

Based on the above displayed model, this study proposes the following hypotheses:

This study aims to provide valuable insights into how AI can be leveraged to improve mental health education among Chinese college students. By understanding the factors that influence students’ engagement with AI technologies, educational institutions can develop more effective strategies to enhance mental health support and academic outcomes. The findings can inform policy decisions, promote the adoption of AI in educational settings, and ultimately contribute to the overall well-being of students. The full set of hypotheses is displayed in Figure 2.

The present study was aimed to test a TPB-based model suggesting that Academic engagement –mental health relationships was serially mediated by students’ attitudes toward AI use, subjective norm, perceived behavioral control and Students’ AI use behavioral intention, which in term influences on mental health among Chinese college students.

The first hypothesis, which predicted a significant positive relationship between students’ engagement and their mental health, was confirmed by the analysis. This finding aligns with existing literature that highlights the importance of academic engagement in promoting students’ mental well-being. For example, Finn and Zimmer (2012) found that higher levels of academic engagement are associated with better mental health outcomes, including reduced anxiety and depression. Similarly, Dumford and Miller (2018) reported that students who are more engaged in their academic pursuits tend to have higher levels of psychological well-being.

However, this finding contrasts with some previous studies that have suggested that the pressure to perform academically can lead to increased stress and mental health issues (Conley et al., 2015; Conley et al., 2017). These studies argue that while engagement can have positive effects, the high expectations and workload associated with academic involvement can also contribute to negative mental health outcomes. Thus, the relationship between engagement and mental health may be more complex than previously thought, necessitating a balanced approach that promotes engagement while also addressing potential stressors.

The second hypothesis posited that attitude toward AI would mediate the relationship between students’ engagement and their mental health. However, the analysis did not find a significant indirect effect through attitude toward AI. This result suggests that while students’ attitudes toward AI are important, they may not play a pivotal role in linking engagement to mental health. This finding is consistent with the other works (Pan, 2020; Zaineldeen et al., 2020), which indicate that attitudes toward technology, while influential, are often overshadowed by more immediate contextual factors such as usability and perceived usefulness. This evidence is in line with a large amount of research showing the predictive power of attitudinal measures on wellbeing (Mason Stephens et al., 2023; Sibley et al., 2020).

Similarly, the findings did not support that subjective norms would mediate the relationship between students’ engagement and their mental health. This result contrasts with studies that emphasize the role of social influence in technology adoption (Venkatesh and Davis, 2000). While subjective norms can affect technology use intentions, their impact on the specific context of mental health outcomes appears limited, highlighting the need for further investigation into other mediating factors. This contrasting evidence can be connected with the TPB-based research about cheating behavior in the academic context (Stone et al., 2010). Given that the use of AI in the classroom, specifically to complete exercises and academic tasks, can be viewed as misconduct, subjective norms can be ambivalent about this behavior. On the one hand, it can be considered a technological advancement and a demonstration of technical capability (Norzelan et al., 2024), but on the other, it can be considered academic misconduct that should be avoided and punished (Maulana, 2019). In this sense, modified versions of the TPB, exploring the justification of this behavior, could clarify the fault of the significance of subjective norm in the present study (Scrimpshire et al., 2017).

The fourth hypothesis, which proposed that perceived behavioral control over AI use would mediate the relationship between students’ engagement and their mental health, was supported. This finding aligns with the Theory of Planned Behavior (Ajzen, 1991), which posits that perceived control is a critical determinant of behavior. Several studies support this notion, particularly in the context of educational technology. For instance, Alalwan et al. (2018) demonstrated that students who feel confident in their ability to use technology, including AI, effectively are more likely to experience positive mental health outcomes, as they can better leverage these tools to enhance their learning and well-being.

However, it is important to recognize that this relationship is not always straightforward. Research by Nazari et al. (2021) highlights that for some students, increased access to AI tools can introduce stress, particularly when they feel overly reliant on the technology or lack full mastery of its use. In such cases, perceived behavioral control may not always lead to positive mental health outcomes, especially if students experience anxiety about their ability to keep up with the demands of using AI. Furthermore, a study by Hsia (2016) emphasizes that perceived behavioral control is a significant predictor of positive outcomes in various digital learning environments. Students who exhibit high levels of perceived control tend to engage more effectively in self-regulated learning, which not only improves academic performance but also contributes to better mental health. In this context, perceived control over AI use can serve as a buffer against stress, enabling students to navigate challenges more effectively.

Thus, while the findings of this study align with the Theory of Planned Behavior and are supported by several studies, it is crucial to consider that the effects of perceived behavioral control over AI use may vary across different student populations. Factors such as digital literacy, confidence in technology use, and individual differences in stress response should be considered in future research to provide a more comprehensive understanding of how AI impacts students’ mental health.

Hypothesis five predicted that AI use behavioral intention would mediate the relationship between students’ engagement and their mental health, and the analysis confirmed this hypothesis. This result supports the idea that behavioral intentions are strong predictors of actual behavior and subsequent outcomes (Sheeran, 2002). When students intend to use AI tools, they are more likely to engage with these technologies in ways that benefit their mental health. This is consistent with findings from Venkatesh et al. (2012), where behavioral intentions regarding technology use have been shown to significantly influence actual engagement with technological tools, ultimately leading to positive educational and psychological outcomes.

Moreover, Hagger et al. (2018) highlight that the translation of intentions into behavior is often facilitated when there is sufficient perceived control, further reinforcing the relationship between AI use intentions and student engagement. In the context of technology adoption, this has been observed in several areas, such as e-learning platforms, mobile applications for educational purposes, and AI-driven tools in education. For instance, in e-learning environments, students’ perceived ease of use and control over the technology significantly impact their intention to engage with online learning tools. Similarly, the adoption of AI-driven platforms like adaptive learning systems, which tailor content based on individual student needs, is strongly influenced by students’ perceived control and confidence in using such technologies (Alqahtani et al., 2023). Therefore, this mediation effect of behavioral intention is robust across various technology contexts, especially in educational technologies where user engagement depends on both the perceived utility of the tool and the user’s ability to navigate it effectively.

Hypothesis six proposed that Chinese college students’ attitude toward AI, subjective norms, and perceived behavioral control over AI use would positively predict their AI use behavioral intention. The analysis confirmed this hypothesis, indicating that these factors collectively shape students’ intentions to use AI. This finding aligns with the Unified Theory of Acceptance and Use of Technology (Venkatesh et al., 2003), which emphasizes the combined influence of attitudes, social norms, and perceived control on technology use intentions.

The predictive power of attitudes, subjective norms, and perceived control on behavioral intentions has been well-documented across several domains. For example, Armitage and Conner (2001), in their meta-analysis, demonstrated that attitudes, followed by subjective norms and perceived control, consistently predicted behavioral intentions in various behavioral contexts, including health-related behaviors. Similarly, Cooke et al. (2016), in their systematic review, confirmed that these constructs strongly influence behavioral intentions, reinforcing their role in shaping actions within technology adoption contexts.

Furthermore, Kumar and Nayak (2023) provided empirical support for the Theory of Planned Behavior, showing that attitudes play a crucial role in influencing behavioral intention, particularly in educational technology. They found that positive attitudes toward AI significantly drive students’ intentions to engage with AI, while subjective norms (influences from peers or faculty) and perceived behavioral control (students’ confidence in using AI) further reinforce this intention. Thus, these findings align with both the Unified Theory of Acceptance and Use of Technology and the broader literature on the Theory of Planned Behavior, reinforcing the notion that attitude is a crucial antecedent of behavioral intention, followed by subjective norms and perceived behavioral control.

Hypothesis seven suggested that AI use of behavioral intention would mediate the relationship between students’ engagement and their mental health. The analysis supported this hypothesis, demonstrating that students who intend to use AI tools are more likely to see improvements in their mental health through increased engagement. This result highlights the importance of fostering strong behavioral intentions to use AI as a pathway to better mental health outcomes. As AI use can be viewed as a resource to deal with highly demanding tasks, behavioral intentions to use them could reduce depression and anxiety associated to academic exhaustion (Li and Anila, 2023).

Finally, hypothesis eight posited that the relationship between students’ engagement and their mental health would be sequentially mediated by their attitude toward AI, subjective norm, perceived behavioral control over AI use, and AI use behavioral intention. The analysis confirmed this complex mediation chain, suggesting that these factors work together in a sequential process to influence mental health outcomes. This finding underscores the interconnected nature of psychological constructs and their collective impact on behavior and well-being.

To sum up, one of the key strengths of this research is its comprehensive approach, examining multiple psychological constructs—attitude, subjective norm, perceived behavioral control, and behavioral intention—in a sequential mediation model. This approach provides a deeper understanding of how these variables interact to influence students’ mental health, moving beyond simple relationships to explore the complexity of behavioral change. By capturing the progression of attitudes, social influences, and self-efficacy beliefs, the study highlights the indirect pathways that link AI engagement with mental health, emphasizing the importance of fostering positive attitudes and enhancing perceived control.

Additionally, the research contributes to the literature by extending the Theory of Planned Behavior (TPB) and Unified Theory of Acceptance and Use of Technology (UTAUT) to AI use in education, demonstrating their relevance in predicting both behavioral and mental health outcomes. The focus on a large sample of Chinese college students enhances the generalizability of the findings, offering timely insights for higher education contexts where AI adoption is rapidly increasing. Methodologically, the use of structural equation modeling (SEM) adds rigor, ensuring that the complex mediation process is empirically validated and the study’s internal validity is strengthened by controlling for confounding variables.

This study provides valuable insights into the mechanisms through which academic engagement influences mental health among Chinese college students, particularly through the lens of AI use. By confirming the direct positive relationship between engagement and mental health, as well as identifying key mediating factors such as perceived behavioral control and AI use behavioral intention, the findings highlight the potential of AI tools in enhancing students’ well-being. However, the non-significant mediating effects of attitude toward AI and subjective norm suggest that more research is needed to fully understand the nuances of these relationships. Overall, the study contributes to the growing body of literature on the role of AI in education and offers practical implications for improving mental health support in academic settings.