In response to the epochal demand for high-quality development in early childhood education in China, it is imperative and necessary to improve the competency level of early childhood educators. The study aims to investigate the relationship between psychological contracts and teacher competency, and to verify the mediating roles of job crafting and professional identity in the relationship between psychological contracts and teacher competency. This study employed validated measurement scales regarding psychological contracts, teachers' professional identity, job crafting, and teacher competency. Each of these scales has established internal consistency coefficients. Data were collected from 318 early childhood teachers in Sichuan, China. The results highlight the significant impact of psychological contracts on the prediction of teacher competency. It is worth mentioning that the psychological contracts, together with their distinct components such as normative responsibility and development responsibility, have a direct and favorable impact on teacher competency. This implies that developing the psychological contracts might be an effective technique for improving teacher competence. The individual mediation of job crafting and professional identity in the link between psychological contracts and teacher competency has been well-established. However, the combined or chain mediating influence of these factors provides a unique and valuable perspective on the phenomenon of job crafting leading to professional identity, which in turn impacts teacher competency. The study found that psychological contracts have a positive predictive effect on teacher competency, while job crafting and professional identity both have independent and chain mediating roles in the relationship between psychological contracts and teacher competency. Therefore, this study suggests a comprehensive enhancement of the psychological contracts level from aspects such as normative responsibility, interpersonal responsibility, and development responsibility. By stimulating job crafting and professional identity levels in both internal and external environments, we can improve the competency level of early childhood educators.

Within the global sustainable development agenda, Sustainable Development Goal 4 (SDG 4) highlights improving the accessibility, quality, and learning experience of technical and vocational education and training (TVET). In China, students in vocational colleges often face greater disparities in academic preparation and access to educational resources than their peers in general higher education. Although artificial intelligence (AI) can provide additional learning support and help mitigate such inequalities, there is little empirical evidence on whether and how Gen-AI usage is associated with vocational students’ learning experiences and emotional outcomes, particularly academic anxiety. This study examines how Gen-AI usage is related to academic anxiety among Chinese vocational college students and explores the roles of class engagement and teacher support in this relationship. Drawing on Conservation of Resources (COR) theory, we analyse survey data from 511 students using structural equation modelling (SEM) and fuzzy-set qualitative comparative analysis (fsQCA). The SEM results indicate that Gen-AI usage is associated with lower academic anxiety, with class engagement mediating this relationship. Teacher support for Gen-AI usage positively moderates the association between Gen-AI usage and class engagement. The fsQCA results further identify several configurations of conditions leading to low academic anxiety. These findings underscore AI’s potential to enhance learning quality and experiences in TVET and provide empirical support for advancing SDG 4 in vocational education contexts.

While the application of artificial intelligence (AI) in education has garnered growing scholarly attention, research specific to vocational education remains limited. In China, vocational education plays a critical role in cultivating skilled workers, yet students in this context often face distinctive academic anxiety and mental health challenges. This study, grounded in the conservation of resources theory, investigates how AI usage influences academic anxiety among Chinese vocational students. Drawing on data from 511 questionnaires, we employed structural equation modeling (SEM) and fuzzy set qualitative comparative analysis (fsQCA) to examine both linear and configurational relationships. The results reveal that AI usage significantly and negatively predicts academic anxiety, with class engagement serving as a key mediating variable. Moreover, teacher support for AI usage, as a conditional resource, positively moderates the link between AI usage and class engagement. The fsQCA further identifies three distinct configurational pathways leading to low academic anxiety. These findings contribute to the literature on AI in vocational education by uncovering its psychological impact and offering theoretical and practical insights for educational management in this domain.

Against the backdrop of sustainable education and the digital transformation of higher education, AI assistants on online learning platforms have become important tools for facilitating students’ access to learning resources, providing immediate feedback, and supporting self-directed learning, thereby creating new opportunities for more effective and sustainable learning experiences. However, the psychological mechanisms linking AI assistant intelligence to students’ innovative behavior remain unclear. Based on Social Cognitive Theory (SCT), this study examines how AI assistant intelligence on online learning platforms relates to higher education students’ innovative behavior through academic self-efficacy and academic procrastination and further explores the moderating role of AI literacy. Using a two-phase time-lagged design, the study collected data through online questionnaires from students at Chinese universities, yielding a final matched sample of 503 students. The findings show that AI assistant intelligence is positively associated with both academic self-efficacy and academic procrastination, but it has no significant direct effect on innovative behavior. Importantly, these two mediators operate in opposing directions: academic self-efficacy transmits a positive indirect effect, whereas academic procrastination transmits a negative indirect effect on innovative behavior. In addition, AI literacy strengthens the positive relationship between AI assistant intelligence and academic self-efficacy while weakening its positive relationship with academic procrastination. These findings reveal the dual psychological pathways through which AI assistant intelligence shapes students’ innovative behavior, extend Social Cognitive Theory in intelligent learning environments, and offer practical implications for leveraging AI-assisted learning to foster students’ innovative capacity and inform the sustainable integration of AI in higher education.