Artificial intelligence (AI) technology has been increasingly integrated into K-12 education, offering a novel pathway for optimizing instructional practices and improving students' academic performance (Wang et al., 2023). This transformative trend is particularly prominent in science education, where AI tools unlock revolutionary potential for teaching complex scientific concepts, delivering personalized learning, and conducting interdisciplinary scientific inquiry (Erduran, 2023). This imposes higher demands on science teachers: they are required to not only possess the professional competence to lead AI-driven instructional innovation, but also acutely perceive and effectively address various ethical issues that emerge continuously in AI-enabled teaching practices.

 In particular, for early career science teachers (ECSTs) in their first three years of appointment, this shift presents both opportunities and challenges, as their core responsibilities entail adapting to technological disruption while establishing foundational professional competencies (Barrett, 2025). ECSTs must not only rapidly master AI-based teaching tools and methodologies to align with evolving instructional requirements, but also navigate the reshaping of professional roles and competency challenges brought about by technological transformation (H et al., 2025). Their AI readiness serves as a pivotal underpinning for addressing the aforementioned pressures.

Against this backdrop, job satisfaction among ECSTs has grown in significance and merits special attention. Neglecting the interactive relationship between AI readiness and job satisfaction will not only directly dampen ECSTs' initiative to proactively integrate AI into teaching and undermine the implementation effectiveness of AI-enabled education, but also adversely affect their career development trajectories. Crucially, low job satisfaction may lead to early career attrition. Surveys indicate that the natural attrition rate of early career teachers (ECTs) has exceeded 30% in some European countries in recent years (OECD, 2021) — a phenomenon that directly reflects their low job satisfaction and thus jeopardizes the stability and sustainable development of the teaching workforce.

 While existing research has examined the association between AI readiness and job satisfaction, it lacks empirical evidence in non-Western contexts. In particular, few studies have systematically explored the mechanism through which four-dimensional AI readiness influences job satisfaction among ECSTs in China. Currently, China's Ministry of Education, Ministry of Science and Technology, and other relevant departments have issued a series of policies to promote and support the full integration of AI into K-12 science education (Ministry of Education et al., 2025). This policy context further underscores the urgency of investigating the practical predicaments faced by this teacher cohort.

 Drawing on the scale developed by Karaca et al. (2021) and the Technology Acceptance Model (TAM), this study adopts the localized framework proposed by Wang et al. (2023), which defines AI readiness as teachers' state of preparedness across the cognitive, ability, vision, and ethical dimensions in the context of AI implementation in educational settings. This study aims to investigate ECSTs' AI readiness and its impact on job satisfaction against the backdrop of AI-enabled science education. Clarifying the inherent correlation between the two is essentially the key to addressing the dilemma of the escalation of technological requirements and the imbalance in teachers' professional adaptation, and it holds great significance for safeguarding the professional growth of ECSTs and laying a solid foundation for the development of the science teaching workforce in the AI era.

This study primarily addressed the following research questions:1.What are the overall level and distribution characteristics of AI readiness among ECSTs?2.What inherent interrelationships exist among the four dimensions (cognition, ability, vision, and ethics) of ECSTs' AI readiness?3.What is the causal relationship between ECSTs' AI readiness and their perceived AI threat, AI-enabled innovation, and job satisfaction?4.Are there significant differences in AI readiness, perceived AI threat, AI-enabled innovation, and job satisfaction among ECSTs with different demographic backgrounds? To address the above questions, convenience sampling was adopted to recruit ECSTs from multiple provinces in China, and a total of 448 valid samples were obtained (189 males and 273 females; average age=25.6 years). The participants' teaching subjects included general science, physics, chemistry, and biology, among others. Data were collected via a questionnaire survey, which was revised based on validated scales (the Teacher AI Readiness Scale and the Job Satisfaction Scale) and with reference to the localized framework proposed by Wang et al. (2023). The questionnaire consisted of 41 items rated on a 5-point Likert scale (1=strongly disagree, 5=strongly agree), and measured seven core variables: four dimensions of AI readiness (cognition, ability, vision, and ethics) adapted from Karaca et al. (2021), perceived AI threat adapted from Mirbabaie et al. (2022), AI-enabled innovation derived from Popenici & Kerr (2017), and job satisfaction derived from Ragu-Nathan et al. (2008). The questionnaire was refined and optimized through back-translation, expert consultation, and a pilot test. Harman's single-factor test was conducted to assess common method bias, and the results showed that the unrotated first common factor explained 37.63% of the total variance (40%). This indicated a low level of common method bias, which did not exert a significant impact on the study results. For data analysis, Partial Least Squares Structural Equation Modeling (PLS-SEM) was performed using the PLSPM package in R software to examine the interrelationships among the variables. Two-step cluster analysis was applied to clarify the distribution characteristics of AI readiness levels, and guided by the Bayesian Information Criterion (BIC), the optimal cluster solution was finally identified. The results of reliability and validity tests demonstrated an ideal level of measurement quality: all Cronbach's α coefficients for the scales were greater than 0.90, and all average variance extracted (AVE) values exceeded 0.77.

This study Reveals the action mechanism of ECSTs' AI readiness and expands the empirical dimensions and scenarios of the Educational Technology Acceptance Theory.The study draws the following conclusions in response to its four core research questions: First, cluster analysis categorized the research sample into three groups: the high AI readiness group (32.59%), the intermediate AI readiness group (49.33%), and the low AI readiness group (18.08%). This finding indicates that nearly half of the ECSTs attain intermediate AI readiness, while a small proportion remain at the low level. Second, regarding the interrelationships among AI readiness dimensions and inter-variable mechanisms, the study confirmed that cognition, ability, and vision all positively shape ECSTs’ awareness of AI ethics, with vision serving as the core driver of ethical practice (the strongest predictive effect). Meanwhile, the impact of AI readiness on ECSTs’ perceived AI threat, AI-enabled instructional innovation, and job satisfaction is focused: ability acts as the key hub, which mitigates ECSTs’ perceived AI threat with a path coefficient of β=-0.35and drives AI-enabled instructional innovation with a path coefficient of β=0.35. Furthermore, AI-enabled instructional innovation is the core positive driver of job satisfaction with a strong path coefficient of β=0.52, whereas perceived AI threat exerts a negative impact on job satisfaction with a path coefficient of β=-0.29. The cognition, vision, and ethics dimensions of AI readiness have no direct effect on the aforementioned outcome variables, with their impacts likely via indirect paths. Finally, in terms of demographic differences, gender exerts heterogeneous effects on the relevant variables: male ECSTs exhibit slightly higher AI readiness than their female counterparts and perform more prominently in AI-enabled innovation, while female ECSTs report a stronger perceived AI threat. This reflects inherent gender differences in technological adaptation and risk perception, thus providing a basis for gender-targeted AI literacy cultivation for ECSTs.

Barrett, A. J. (2025). Immersion in virtual reality-based teacher training simulations with artificial intelligence-integrated student agents to increase learning achievement for early-career teachers. (Order No. 32045788, The Florida State University).ProQuest Dissertations and Theses,,108. Erduran Sibel.(2023).AI is transforming how science is done. Science education must reflect this change..Science (New York, N.Y.),382(6677),eadm9788-eadm9788. H, R. E., & Efe, A. (2025). Science Teachers’ Perceptions Of The Artificial Intelligence In Science Education: Challenges, Readiness, Benefits, And Impact On Student Learning.Journal of Baltic Science Education,24(4), 655-669. Karaca, O., Çalıs¸kan, S. A., & Demir, K. (2021). Medical artificial intelligence readiness scale for medical students (MAIRS-MS)–development, validity and reliability study. BMC Medical Education, 21(1), 1–9. Mirbabaie, M., Brünker, F., Mollmann ¨ Frick, N. R., & Stieglitz, S. (2022). The rise of artificial intelligence–understanding the AI identity threat at the workplace. Electronic Markets, 32(1), 73–99. Popenici, S. A., & Kerr, S. (2017). Exploring the impact of artificial intelligence on teaching and learning in higher education. Research and Practice in Technology Enhanced Learning, 12(1), 1–13. Ragu-Nathan, T., Tarafdar, M., Ragu-Nathan, B. S., & Tu, Q. (2008). The consequences of technostress for end users in organizations: Conceptual development and empirical validation. Information Systems Research, 19(4), 417–433. Wang Xinghua,Li Linlin,Tan Seng Chee,Yang Lu &Lei Jun.(2023).Preparing for AI-enhanced education: Conceptualizing and empirically examining teachers’ AI readiness.Computers in Human Behavior,146. Wang, N., & Lester, J. (2023). K-12 education in the age of AI: A call to action for K-12 AI literacy.International Journal of Artificial Intelligence in Education,33(2), 228-232.