Understanding AI ethics in education with a LLM-based chatbot: Evidence from PLS-SEM and fsQCA analyses

Understanding AI ethics in education with a LLM-based chatbot: Evidence from PLS-SEM and fsQCA analyses | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Understanding AI ethics in education with a LLM-based chatbot: Evidence from PLS-SEM and fsQCA analyses Chenchen Liu, Yangying Yi, Xinying Wang, Haijie Wang, Youmei Wang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8618094/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 11 You are reading this latest preprint version Abstract As Artificial Intelligence (AI) becomes increasingly and widely integrated in education, fostering teachers’ understanding of AI ethics has become critical. This study investigates the acceptance of a chatbot based on a Large Language Model (LLM), designed to enhance teachers’ understanding of AI ethics in education. A mixed-methods approach combining Partial Least Squares Structural Equation Modeling (PLS-SEM) and fuzzy-set Qualitative Comparative Analysis (fsQCA) was applied in the current study with 142 participants. Findings show that the content accuracy, system functionality and stability significantly influence perceived ease of use and perceived usefulness. FsQCA further revealed five sufficient configurations associated with high continuance intention. These findings provide practical guidance for designing AI-supported learning environments and offer directions for future research. Social science/Education Business and commerce/Information systems and information technology Physical sciences/Mathematics and computing Social science/Science technology and society AI ethics PLS-SEM fsQCA Technology acceptance model Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction With the continuous innovation of technology, Generative Artificial Intelligence (GenAI) is becoming increasingly and widely used in the field of education, and its rapid integration in educational environments is profoundly changing the way of teaching and learning. While GenAI offers significant opportunities for enhancing teaching efficiency and personalized instruction (Bozkurt, 2023 ; Wang et al., 2023 ), it also raises a range of key ethical questions about transparency, fairness, accountability, and changing roles of educators (Jobin et al., 2019 ; Kazim & Koshiyama, 2021 ; Nguyen et al., 2023 ). Research suggests that over-reliance on AI can impair the ability of teachers and students to think independently and creativity, leading to academic misconduct such as cheating, plagiarism, or using AI to complete assignments and write essays (Bakar-Corez & Kocaman-Karoglu, 2024 ; Kasneci et al., 2023 ). Additionally, AI-generated answers and content can hinder critical thinking for teachers and students, bypassing traditional learning processes, thereby damaging their academic development and the credibility of the education system (Ipek et al., 2023 ). Additionally, the information generated by GenAI tools may be inaccurate or biased, further increasing the risk of academic dishonesty and copyright infringement (Baek & Kim, 2023 ; Eke, 2023 ). In response to the growing ethical issue of AI, studies have pointed out that a complete rejection of AI is not a viable solution (Lim et al., 2023 ). Many scholars and educational institutions advocate for an open and cautious approach, emphasizing the development of clear frameworks and clear guiding principles to harness the full potential of AI while effectively mitigating its potential risks (Uddin, 2024 ). While numerous ethical guidelines and policies for AI have emerged in recent years, these norms often remain abstract and lack concrete implementation guidance. Thus, teachers struggle to clearly identify and respond to the ethical risks and responsibilities of using AI in educational contexts (Nikolopoulou, 2021 ). At the same time, teachers play a vital role in guiding students to use AI responsibly and consciously, their ethical awareness and literacy are essential for the sustainable integration of AI in education (Ayanwale et al., 2024 ). However, the opportunities for teachers to participate in AI ethics learning are still limited and somehow ineffective. The lack of structured professional development pathways limits educators’ ability to meaningfully engage in and contribute to discussions and practices related to AI ethics (Lim et al., 2023 ). This suggests that existing policies and guidelines alone are insufficient to address current challenges, highlighting the need for targeted training mechanisms to systematically enhance teachers’ cognition, understanding, and practical competence in AI ethics (Mouta et al., 2025 ). Conventional training approaches frequently prove inadequate in addressing the specific demands of this domain. On the one hand, conventional training is frequently detached from practical teaching contexts and lacks situational relevance. On the other hand, its content often lacks flexibility and continuity, limiting its effectiveness in keeping pace with fast-evolving technological developments.(Hagendorff, 2020 ; Wiese et al., 2025 ). Accordingly, it is imperative to design innovative and scalable training models that facilitate teachers’ continuous professional development and promote deep, context-aware ethical reflection. To address this need, this study introduces chatbot, a large language model-based chatbot developed via the LangChain framework, designed to support teachers in enhancing their ethical literacy in the context of artificial intelligence. To understand teachers’ intention to use the chatbot and the factors influencing it, this study adopts the Technology Acceptance Model (TAM) to examine how key variables—such as perceived usefulness and perceived ease of use—affect their usage intentions. TAM has been widely substantiated in many fields to effectively explain user behavior in adopting information systems, particularly focusing on perceived usefulness and perceived ease of use as key determinants factors affecting user intention and actual usage behavior (Barrett et al., 2021 ; Davis, 1989; Wang et al., 2021 ). But as research advances further, more and more scholars have begun to pay attention to other potential influencing factors other than the TAM model (Barrett et al., 2021 ; Kim & Lee, 2014 ). Studies have shown that system quality is also an important factor affecting user adoption behavioral intention to adopt technology (Shen & Liu, 2022 ; Xu et al., 2013 ; Zheng et al., 2013 ), high-quality systems often enhance user trust, satisfaction, and overall user experience, thereby increasing their willingness to adopt and continue using them (Almaiah & Mulhem, 2019; Al-Adwan et al., 2022 ). This study introduces a specific dimension of AI ethics education based on TAM, focusing on three key factors affecting the system quality of the chatbot: content accuracy, system functionality and stability, and technical support. These factors are hypothesized to affect users’ perceptions of usefulness and ease of use, further affecting their behavioral intentions to continue using the chatbot. By integrating PLS-SEM and fsQCA, this study comprehensively reveals the causal relationships and multiple configuration paths that affect teachers’ willingness to use AI ethics learning tools, providing a new theoretical perspective for understanding their motivations. Literature review Technology acceptance model The Technology Acceptance Model (TAM) was first introduced by Davis (1989) to explore the degree of user acceptance of emerging technologies. As a maturing theoretical model, TAM has been widely used to understand the key factors influencing user adoption and sustained use of technological innovations (Al-Emran et al., 2018; Teo, 2010). Over the past decades, it has been successfully applied in various fields, including mobile learning (Almaiah et al., 2016; Gómez-Ramirez et al., 2019), smart learning environments (Hu, 2022), virtual learning platforms (Al-Adwan et al., 2023), and online education systems (Han & Sa, 2022; Waheed & Jam, 2010). With the expansion of its application contexts, an increasing number of studies have explored the applicability of the TAM model in the context of diversification, and proposed extensions to its original structure. The model emphasizes two key constructs: Perceived usefulness (PU) and perceived ease of use (PEU), which are considered crucial in determining users’ attitudes and behavioral intentions toward a particular technology. Perceived usefulness refers to the degree to which users believe that a technology can improve their performance, while perceived ease of use refers to the degree to which users find the technology free from effort (Davis, 1989; Liu et al., 2010). These perceptions directly influence users’ intention to use the technology, which serves as the primary outcome variable in TAM. It is widely regarded as a strong predictor of actual usage behavior and is critical in assessing the effectiveness of educational technology adoption (Barrett et al., 2021; Scherer et al., 2019; Zhang et al., 2023). Numerous studies have consistently shown that among the key constructs of the TAM, PU is typically a strong predictor of teachers’ intentions to adopt new technologies. Meta-analyses by Ma and Liu (2004) and King and He (2006) show that PU has a wide range of coefficient values, indicating its strong impact on users’ intentions to use various technological tools. This effect is particularly evident in the educational context, where PU serves as a crucial determinant of teachers’ intentions to adopt new technologies. When teachers perceive that a technology can enhance instructional effectiveness, streamline teaching processes, or facilitate student learning, they are more likely to adopt it in their teaching practices (Bai et al., 2021; Scherer et al., 2015). Similarly, complementing the role of PU, PEU is likewise supported by educational technology research as a key predictor of teachers’ ITU. For instance, Teo (2010) found that PEU had a significant positive influence on pre-service teachers’ intention to use educational software, indicating that ease of system interaction is an important motivational factor. Weng et al. (2018) also confirmed in their TAM-based study on elementary school teachers’ use of multimedia materials that PEU not only has a direct effect on ITU, but also exerts an indirect effect through attitude toward use. These findings provide the basis for the following two hypotheses of this study： H1. PEU will have a positive and significant effect on teachers’ ITU of the chatbot. H2. PU will have a positive and significant effect on teachers’ ITU of the chatbot. The quality of learning system The Information System Success Model (IS Success Model), proposed by DeLone and McLean (1992), is one of the most widely adopted frameworks for evaluating the effectiveness of information systems. This model conceptualizes system success through three primary dimensions: information quality, system quality, and service quality. Subsequently, Rai et al. (2002) provided empirical validation demonstrating that the three dimensions of the information system success model are key constructs and factors influencing IS usage and performance. This study modifies the original IS success model to align more closely with the educational application scenario of the AI ethics learning tool. Specifically, information quality is revised as “content accuracy”, referring to the relevance and correctness of the ethical content generated by the chatbot; system quality is reformulated as “system functionality and stability”, capturing the chatbot’s operational reliability and technical performance; and service quality is renamed “technical support”, assessing the quality and responsiveness of assistance provided to users. This study draws on Cheng’s (2012) empirical findings that information quality, system quality, and service quality are the key quality antecedent variables influencing users’ acceptance of e-learning systems. Based on the research conclusions of Cheng (2012), this study defines these three quality dimensions as external antecedents affecting teachers’ PU and PEU. In the proposed research model, these quality dimensions function as external variables that influence teachers’ system perceptions and ultimately shape their intention to adopt the AI-based ethical learning tool. Content Accuracy Content accuracy refers to the ability of a system to provide correct, trustworthy, and up-to-date information (DeLone & McLean, 2003; Sassano et al., 2010). As a key component of information quality, content accuracy is fundamental in shaping users’ trust and overall evaluation of a technology system, particularly in knowledge-intensive environments (Nelson et al., 2005). In AI-supported tools, maintaining high content accuracy is particularly important to ensure teaching reliability and professional credibility as content is dynamically generated (Almuhanna, 2024). Seddon (1997) points out that the user’s perception of the usefulness of a system is greatly influenced by the quality of the information it provides. Machdar (2016) further emphasized that when users rely on system-generated content for decision-making, higher information quality makes the system more likely to be perceived as useful and trustworthy. For teachers, the content accuracy of AI-generated tools significantly enhances their confidence in applying these tools to teaching tasks, such as writing assistants or instructional simulations (Aljuaid, 2024; Kim & Kim, 2022). In addition, studies have found that teachers are more inclined to adopt systems that provide accurate and highly relevant resources in AI-based intelligent recommendation systems (Siafis et al., 2024). Beyond fostering users’ perceived usefulness, content accuracy also facilitates perceived ease of use. When information is accurate and clearly presented, users can more easily understand how to operate the system, thereby enhancing their perception of its ease of use (Almaiah & Al Mulhem, 2019). For example, in digital assessment systems, when teachers receive accurate and well-structured feedback, they are less likely to need additional explanation or manual correction, which reduces disruptions during use and enhances the system’s operability (Patra et al., 2022). Moreover, in language education, teachers generally believe that high content accuracy as a key factor in reducing the need for repeated verification of system outputs. This reduction in operational pressure and technical anxiety further strengthens their perception of system ease of use (Belda-Medina & Calvo-Ferrer, 2022; Hsu, 2016). Based on these insights, the following hypotheses are proposed: H3. CA of the chatbot will be a positive and significant predictor of PU. H4. CA of the chatbot will be a positive and significant predictor of PEU. System functionality and stability System functionality and stability refer to the overall technical performance demonstrated by the system during its operation, including operational stability, functional reliability, system response speed, interface usability, and cross-platform compatibility (Almaiah & Alamri, 2018; DeLone & McLean, 2003; Giang & Nga, 2024; Kim & Lee, 2014; Ohliati & Abbas, 2019). This variable examines the system’s ability to provide stable, uninterrupted service, ensure smooth access to its functions, and support users in completing tasks efficiently. System functionality and stability are key characteristics of a information system that can significantly influence users’ attitudes and behaviors towards its use (Almaiah & Al Mulhem, 2019). A system with comprehensive features and high stability can provide a more efficient and smoother user experience, fostering users’ belief that it can effectively meet their needs and thereby enhancing its perceived usefulness (Kim & Lee, 2014). According to Farhan et al. (2019), system quality dimensions—including reliability, response time, and interface design—have a significant impact on teachers’ perceived usefulness of school e-learning systems. When a system is highly functional and stable, teachers are more likely to trust the technology and believe in its potential to improve teaching organization, classroom interaction, and student engagement. As Du et al. (2022) point out, when evaluating the usefulness of virtual reality technology in classroom teaching, teachers pay particular attention to whether the system can maintain continuous and stable operation and whether it can support diverse teaching situations. System functionality and stability play a critical role in shaping users’ PEU. A successful online learning system should be evaluated in terms of user-friendliness and the effectiveness of feedback mechanisms to ensure that learners can operate smoothly and receive timely support (Giang & Nga, 2024; Mahande et al., 2019). For teachers, A reliable and user-friendly system allows them to navigate it easily, build teaching confidence, and minimize frustration (Ghavifekr & Rosdy, 2015; Padayachee, 2017). Building on this, Ke et al. (2012) demonstrated that systems with well-designed interfaces and stable performance can help teachers master and use the technology to teach more efficiently and increase student participation in the classroom. Based on these insights, the following hypotheses are proposed: H5. SFS of the chatbot will positively and significantly predict PU. H6. SFS of the chatbot will positively and significantly predict PEU. Technical Support Technical support refers to the level of service provided by the information system to assist users. It includes the timeliness of responses, the effectiveness of problem resolution, the professionalism of support staff, and the adequacy of user training and guidance (Dwivedi et al., 2011; Grönroos, 1984; Ting et al., 2011). During system usage, users may encounter system errors, difficulties in operating specific functions, or incompatibility due to system updates. Timely intervention by technical support can significantly reduce barriers to use and enhance user trust and satisfaction with the system (Gajic & Boolaky, 2015; Hu et al., 2009; Kalankesh et al., 2020). Technical support has a positive impact on PU by improving the reliability and operation guarantee of the system and enhancing the use’s trust in the actual utility of the system. In educational contexts, teachers are often not technical experts. When faced with unfamiliar teaching tools, timely and professional technical assistance can help them better perceive the system’s support for their teaching objectives (Ghavifekr & Rosdy, 2015; Padayachee, 2017). In addition, empirical studies on teachers’ use of virtual reality technology in the classroom have shown that technical support has a significant impact on the perceived usefulness of their systems, which in turn further affects teachers’ user satisfaction (Du et al., 2022). This further shows that technical support, as a core element of service quality, is a key factor in shaping the perception of teachers’ system usefulness. At the same time, technical support also plays an important role in the user’s PEU. Timely and effective assistance can reduce cognitive and technical barriers during operation, allowing teachers to use system functions more smoothly (Arteaga Sánchez et al., 2013; Ibrahim & Shiring, 2022). Empirical research further confirms this view. In a study on the use of WebCT teaching systems by college teachers, Kim and Lee (2014) found that the continuous technical support of the chatbot significantly reduced teachers' uncertainty in the face of technology, thereby improving their overall perception of the ease of use of the system. In addition, Dwivedi et al. (2011) pointed out that technical support improves system usability by providing assistance in key usage links, thereby enhancing the user’s perception of system learnability and intuitiveness. Based on these insights, the following hypotheses are proposed: H7.TS of the chatbot will be a positive and significant predictor of PU. H8.TS of the chatbot will be a positive and significant predictor of PEU. Theoretical confirmatory model Based on the synthesis of prior studies outlined in Subsections 2.1 and 2.2, this study formulates a confirmatory theoretical framework (see Figure 1) to empirically explore teachers’ acceptance of the chatbot. In the proposed model, content accuracy, system functionality and stability, and technical support are included as external variables that influence teachers’ perceived ease of use and perceived usefulness, which in turn affect their intention to use the chatbot. This model provides a structured framework for understanding key factors that drive technology adoption. Research methods Sample and data collection The research employed a stratified random sampling method, selecting, teachers from a public university located in southeastern China as the research subjects. A questionnaire was formulated after referring to previous studies and consulting with experts to ensure the validity of the questionnaire. As depicted in Fig. 1 , the three structures are named “Content accuracy,” “System function and stability”, and “Technical support.” All items were evaluated using a five-point Likert scale, with a score range from 1 (strongly disagree) to 5 (strongly agree). To ensure the reliability and validity of the collected data, the research team set a critical criterion: all teachers who participated in the survey must have had at least one month of experience using the chatbot. Figure 2 shows a participant interacting with the chatbot. This criterion not only helps in obtaining feedback based on actual usage experiences but also reduces cognitive biases due to short usage times, ensuring that the research results accurately reflect teachers’ technical acceptance and usage intentions of the chatbot. Under the strict adherence to the above sampling principles and standards, a total of 168 questionnaires were initially collected in this study. After a meticulous screening process, including the elimination of questionnaires with blank filling, missing key information, and arbitrary filling, 142 valid questionnaires were finally retained for further analysis. Chatbot This study presents the design of an AI-assisted chatbot for ethical learning and reflection, which integrates a large language model (LLM) through API access. The chatbot leverages the LLM’s natural language processing capabilities to facilitate structured dialogue and critical engagement with ethical issues in educational contexts, aiming to support the development of teachers’ ethical awareness and reasoning. The chatbot is designed around three main functions: knowledge expansion, case study, and post-learning evaluation to support the acquisition, application, and evaluation of ethical knowledge. Through the knowledge expansion module, teachers can access standardized explanations of fundamental concepts, principles, and common issues related to AI ethics by interacting with a large language model, thereby enhancing their theoretical understanding. The case-based learning module offers a rich repository of educational ethics cases, enabling teachers to engage with concrete scenarios, analyze ethical dilemmas, and internalize ethical knowledge by applying theory to practice. The post-learning assessment module dynamically generates evaluation items to assess teachers’ mastery of ethical knowledge, providing personalized feedback that fosters ongoing reflection and professional growth. Below is a detailed explanation of the functions of each module (see Fig. 3 ). Data analysis This study employed SmartPLS 4.0 for Partial Least Squares Structural Equation Modeling (PLS-SEM), complemented by fuzzy-set Qualitative Comparative Analysis (fsQCA), to comprehensively analyze the factors influencing teachers’ intention to continue using the chatbot. PLS-SEM was selected for its robust handling of small sample sizes (n = 142) and its ability to estimate complex structural models with multiple latent constructs, such as content accuracy, system functionality and stability, technical support, perceived usefulness, perceived ease of use, and intention to use. This method is suitable for exploratory research and theoretical development with low distribution assumptions (Hair et al., 2021). To enrich the findings and address the complexity of configurations in the user decision-making process, the fsQCA method is also introduced. PLS-SEM is primarily used to test linear net effects between constructs, while fsQCA is able to identify multiple causal paths that lead to the outcome variable—intention to use—through a combination of conditions such as high content accuracy, functional stability, and ease of use. This combined strategy provides both variance-based and set theory-based insights. Data analysis is divided into three steps: (1) Measurement model assessment, involving the evaluation of indicator loadings, composite reliability, and average variance extracted to ensure the validity and reliability of latent constructs. (2) Structural model assessment, including estimation of path coefficients and explanatory power (R²) to test hypothesized relationships among constructs. (3) FsQCA, which was used to identify sufficient configurations of causal conditions that lead to high continuance intention, thus capturing the multifaceted and equifinal nature of user acceptance in ethical AI learning tools. Results Measurement model estimation The reliability and validity of the latent variables were tested based on previous studies evaluating the measurement model (Hair et al., 2017 ), as shown in Table 1 . All individual projects have a reliability (load) of 0.635 or higher, meeting the criteria for reliability of individual projects (Hajiar, 2018). Internal consistency was assessed using both composite reliability (CR) and Cronbach’s alpha. In this study, the Cronbach’s alpha ranged from 0.833 to 0.914, exceeding the commonly accepted threshold of 0.7, thereby indicating satisfactory reliability (Onwuegbuzie & Daniel, 2002 ). Additionally, the CR values for all constructs fell between 0.885 and 0.933, which is well above the minimum recommended level of 0.70 and within the ideal range of 0.70 to 0.90, reflecting strong internal consistency. Convergent validity (CV) was examined through the Average Variance Extracted (AVE). The AVE scores in this study varied between 0.523 and 0.768, all surpassing the benchmark value of 0.50 as proposed by Chin ( 1998 ) and Fornell and Larcker ( 1981 ). These results suggest that the measurement model demonstrates adequate convergent validity. Discriminative validity refers to the ability of a variable to be distinguishable from other variables to some extent. The Fornell-Larcker criterion is a method used to test discriminative validity. According to this criterion, the condition for determining validity is that the square root of the mean variance extraction for each variable should be greater than the correlation coefficient between that variable and any other variable (Fornell & Larcker, 1981 ). The results in Table 2 show that the square root of the average variance extracted for each variable is greater than its highest correlation coefficient with other variables, thus meeting the requirements for discriminative validity. In conclusion, this study systematically tests the reliability and validity of the measurement model, and the results show that the factor load of each variable exceeds 0.7, and the internal consistency and aggregate validity meet the standard. This indicates that the measurement model has high reliability and validity, and is suitable for subsequent structural model analysis. Table 1 CFA results for reliability and validity of measurement items Construct Original items Factor loadings Cronbach’s alpha Composite reliability Average variance extracted (AVE) Content accuracy CA1 0.743 0.871 0.907 0.66 CA2 0.833 CA3 0 842 CA4 0.781 CA5 0.859 System functionality and stability SFA1 0.716 0.848 0.885 0.523 SFA2 0.725 SFA3 0.752 SFA4 0.751 SFA5 0.721 SFA6 0.667 SFA7 0.727 Technical support TS1 0.758 0.833 0.889 0.667 TS2 0.827 TS3 0.852 TS4 0.827 Perceived ease of use EU1 0 635 0.892 0.916 0.612 EU2 0.832 EU3 0.836 EU4 0.829 EU5 0.834 EU6 0.788 EU7 0.697 Perceived usefulness U1 0.847 0.914 0.933 0.7 U2 0.803 U3 0.815 U4 0.878 U5 0.789 U6 0.885 Intention to use WCU1 0.865 0.899 0.93 0.768 WCU2 0.891 WCU3 0.875 WCU4 0.875 Table 2 Discriminant validity results Content accuracy Perceived ease of use System functionality and stability Technical support Perceived usefulness Intention to use Content accuracy 0.813 Perceived ease of use 0.633 0.782 System functionality and stability 0.644 0.67 0.723 Technical support 0.404 0.497 0.509 0.817 Perceived usefulness 0.606 0.871 0.661 0.457 0.837 Intention to use 0.653 0.67 0.648 0.397 0.716 0.876 Structural model estimation PLS-SEM analyzes a structural model by the indicators of: (1)R 2 (R-squared/coefficient of determination); (2) Q 2 (The predictive relevance); (3)path coefficient effect size; (4) SRMR (Standardized Mean Root Square Residual); (5) VIF(Variance Inflation Factor); (6) BIC (Bayesian Information Criterion). First, potential multicollinearity in the structural model was examined by calculating the Variance inflation factor (VIF) values, and it was found that all values were below 3.30 (Table 3 ), well below the recommended threshold of < 5 (Hair et al., 2017 ). These results confirm that multicollinearity has not reached a critical level among the predictor constructs in the model. Furthermore, the cross-validated redundancy (Q 2 ) is greater than zero, indicating that the model has predictive relevance. The direct path coefficients of the structural model were analyzed to examine the hypothesized relationships between variables. Based on a sample of 142 respondents, the following findings were observed. The analysis of the coefficients in the TAM constructed in Fig. 1 revealed that the path from CA to PEU was small but statistically significant ( β = 0.321, p < 0.01), supporting H3. The path from SFS to PEU was also significant ( β = 0.374, p < 0.01), supporting H5. In addition, the path from TS to PEU was significant as well ( β = 0.176, p < 0.05), providing support for H7. Regarding PU, the path from CA to PU was significant ( β = 0.291, p < 0.01), supporting H4, and the path from SFS to PU was also significant ( β = 0.406, p 0.05), leading to the rejection of H8. As for ITU, the path from PEU to ITU showed a small positive effect but was not significant ( β = 0.191, p > 0.05), so H1 is not supported. In contrast, the path from PU to ITU showed a strong and statistically significant positive effect ( β = 0.550, p < 0.001), thus H2 is supported. For a complete representation of the loadings, see Fig. 4 and Table 4 . Table 3 Saturated model results R 2 Adj. R 2 Q 2 RMSE MAE SRMR VIF BIC Chi-square PEU 0.541 0.531 0.488 0.731 0.52 0.071 < 3.30 -91.845 1066.211 PU 0.506 0.495 0.439 0.772 0.536 -81.213 ITU 0.522 0.515 0.441 0.762 0.599 -90.826 Table 4 Path coefficients and hypothesis testing Hypothesis Relationships Mean Standard deviation t-statistics p-value Results H1 PEU -> ITU 0.189 0.137 1.382 0.167 Unsupported H2 PU -> ITU 0.552 0.132 4.185 0.000 Supported H3 CA ->PEU 0.318 0.114 2.977 0.003 Supported H4 CA -> PU 0.296 0.117 2.722 0.007 Supported H5 SFS -> PEU 0.382 0.132 2.772 0.006 Supported H6 SFS -> PU 0.404 0.141 2.707 0.007 Supported H7 TS -> PEU 0.178 0.077 2.225 0.026 Supported H8 TS -> PU 0.133 0.072 1.815 0.07 Unsupported Qualitative Comparative Analysis of Fuzzy Sets (fsQCA) Selection and calibration of variables Based on the literature and empirical research findings, CA, SFS, TS, PEU, and PU are selected as condition variables, while ITU serves as the outcome variable. Following the approach of Rihoux and Ragin ( 2009 ), the 95th percentile, 50th percentile, and 5th percentile of the data are set as calibration anchors, representing full membership, the crossover point, and full non-membership, respectively. The calibration process is carried out using fsQCA 4.1 software. Necessity analysis of conditions Before conducting the configurational analysis, the necessity of each condition variable (including its negation) is tested to determine whether it constitutes a necessary condition for ITU. This is done through an analysis of the necessity of each individual condition variable for the outcome variable (as shown in the Table 5 ). The results indicate that the consistency of all condition variables is less than 0.9, meaning that none of them are necessary conditions for a high/low ITU. Therefore, further configurational analysis is required. Table 5 Details of the necessity analyses Condition Variables ITU ~ITU Consistency Coverage Consistency Coverage CA 0.826 0.831 0.649 0.587 ~CA 0.590 0.590 0.814 0.808 SFS 0.878 0.799 0.694 0.568 ~SFS 0.525 0.656 0.754 0.847 TS 0.773 0.763 0.653 0.580 ~TS 0.574 0.648 0.733 0.744 PEU 0.809 0.844 0.621 0.583 ~PEU 0.600 0.638 0.834 0.797 PU 0.858 0.833 0.675 0.589 ~PU 0.576 0.663 0.808 0.837 Note: “~”denotes the set relationship of “negation” or “complement” Sufficiency analysis of condition configurations Following the approach of Pappas & Woodside ( 2021 ) and Leppänen (2023), for medium-sized samples (50 < N < 150), the frequency threshold is set at 2, and the raw consistency threshold is set at 0.8. After constructing the truth table, Standard Analyses are performed. Given the numerous factors influencing teachers’ ITU in the research context, and the difficulty in measuring the necessity of each factor, the condition variables are uniformly classified as “Present or Absent”. In QCA studies, intermediate solutions, which are reasonable, well-founded, and of moderate complexity, are typically the preferred choice for reporting and interpretation. Therefore, the analysis primarily focuses on the intermediate solution, with simple solutions used as supplementary, ultimately identifying the core and peripheral conditions influencing the outcome variable. Under the Intermediate Solution type, eight configurational paths are identified, but the results are not concise. After adjusting the frequency threshold to 3 and performing Standard Analyses again, five configurational paths are obtained. The reduction in the number of paths decreases redundancy between them, making the analysis clearer. While the Solution Coverage of these five paths slightly decreases (from 0.878 to 0.836), it remains at a high level; however, Solution Consistency improves (from 0.788 to 0.850), surpassing the acceptable minimum threshold of 0.8, indicating that the logic of the paths is more robust. The adjustment of the frequency threshold is thus deemed appropriate. Therefore, the five configurational paths presented in Table 6 can be regarded as the sufficient conditions for influencing teachers' intention to continue using the system in the given research context. Table 6 Sufficiency Analysis of Condition Configurations Condition Variables ITU 1 2 3 4 5 CA • × × × SFS ⊙ ⊙ ⊙ × × TS ⊙ × ⊙ PEU ⊙ × ⊙ × PU • × × • Consistency 0.918 0.907 0.860 0.915 0.937 Raw coverage 0.664 0.744 0.398 0.294 0.325 Unique coverage 0.026 0.084 0.023 0.009 0.002 Solution consistency 0.850 solution coverage 0.836 Note: “•” denotes peripheral conditions, “×” denotes non−existent conditions (negated conditions), “⊙” denotes core conditions, and a blank space indicates conditions that are optional . Configuration 1 (CA*SFS*TS) indicates that, in the process of teachers using the artificial intelligence ethics learning chatbot, CA serves as a peripheral condition providing support, while SFS and TS act as core conditions playing a critical role. This suggests that when the system functions are stable and technical support is adequate, the accuracy of the content enhances teachers’ trust in the chatbot and their intention to use it. Therefore, Configuration 1 can be labeled as “Technological quality satisfaction type”. The consistency of this configuration is 0.918, with a raw coverage of 0.664 and unique coverage of 0.026. This means that this combinational path can explain 66.4% of the cases related to teachers’ intention to accept the chatbot, and 2.6% of the cases can be exclusively explained by this path. Configuration 2 (SFS*PEU*PU) suggests that, during the use of the artificial intelligence ethics learning chatbot, SFS and PEU play core roles, while PU serves as a peripheral condition offering supplementary support. This implies that when the chatbot’s system functionality is stable and teachers find the chatbot easy to use, although perceived usefulness is not a decisive factor, it still helps to some extent in enhancing teachers’ intention to use the chatbot. Thus, Configuration 2 can be named “System experience satisfaction type”. The consistency of this configuration is 0.906, with a raw coverage of 0.744 and unique coverage of 0.084. This indicates that the combinational path can explain 74.4% of the cases regarding teachers’ intention to accept the chatbot, and 8.4% of the cases can be independently explained by this path. Configuration 3 (~ CA*SFS*~PEU*~PU) shows that, in the process of using the artificial intelligence ethics learning chatbot, SFS serves as the sole core condition playing a crucial role. This suggests that when the chatbot’s system functionality is stable, the absence of other conditions does not significantly affect teachers’ intention to use the chatbot. The stability of the system itself provides teachers with confidence and assurance to continue using the chatbot. Therefore, Configuration 3 can be labeled as “Functionality stability satisfaction type”. The consistency of this configuration is 0.860, with a raw coverage of 0.397 and unique coverage of 0.023. This means that this combinational path can explain 39.7% of the cases related to teachers’ intention to accept the chatbot, and 2.3% of the cases can be independently explained by this path. Configuration 4 (~ CA*~SFS*~TS*PEU*~PU) indicates that, in the process of using the chatbot, PEU serves as the sole core condition playing a key role. This suggests that when teachers perceive the chatbot as easy to operate, the absence of other conditions does not significantly influence their intention to use it. The chatbot’s ease of use alone is sufficient to motivate teachers to continue using it. Thus, Configuration 4 can be labeled as “Operational convenience satisfaction type”. The consistency of this configuration is 0.914, with a raw coverage of 0.293 and unique coverage of 0.008. This implies that this combinational path can explain 29.3% of the cases related to teachers’ intention to accept the chatbot, and 0.8% of the cases can be exclusively explained by this path. Configuration 5 (~ CA*~SFS*TS*~PEU*PU) suggests that, during the use of the chatbot, SFS acts as a core condition, while PU serves as a peripheral condition providing auxiliary support. This implies that when the system functionality is stable, teachers’ perceived usefulness of the chatbot will further enhance their intention to use it. Therefore, Configuration 5 can be labeled as “System value satisfaction type”. The consistency of this configuration is 0.936, with a raw coverage of 0.325 and unique coverage of 0.002. This indicates that the combinational path can explain 32.5% of the cases related to teachers’ intention to accept the chatbot, and 0.2% of the cases can be explained by this path. Conclusion, implications and limitations This study integrates the TAM with a mixed-methods approach, combining PLS-SEM and fsQCA, to uncover the diverse factors influencing teachers’ intention to continue using the chatbot. Based on data from 142 participants, PLS-SEM results indicate that perceived usefulness significantly and positively impacts behavioral intention, while perceived ease of use does not show a significant direct effect. Further fsQCA analysis identifies five sufficient configurations leading to high intention to continue use: Technological quality satisfaction, system experience satisfaction, functionality stability satisfaction, operational convenience satisfaction, and system value satisfaction. These findings highlight the multidimensional nature of technology acceptance in AI ethics education and offer empirical evidence and practical implications for the design of effective and AI-supported learning tools. The results of the PLS-SEM analysis showed that perceived usefulness had a significant positive impact on teachers’ behavioral intention to use the chatbot. This finding is consistent with previous research on continued use intention of information systems (Daneji et al., 2019 ; Huang, 2019 ). Foroughi et al. (2023) further pointed out that in educational contexts, perceived usefulness is an important driving force for teachers to adopt AI tools, especially when the chatbot can effectively support ethical reasoning and teaching practices. The results indicate that when teachers perceive the chatbot’s value in improving their understanding and application of AI ethics, they are more likely to continue using it. In contrast, perceived ease of use had a nonsignificant direct effect on intention to use, which may be attributed to teachers’ high overall digital literacy and therefore their lower sensitivity to usability barriers (Amhag et al., 2019 ; Saxena & Doleck, 2023). Ngo et al. ( 2024 ) further point out that in scenarios with high cognitive load, functional value is often more influential than ease of use in technology adoption. In terms of external factors, content accuracy and system function significantly improve perceived usefulness and perceived ease of use, indicating that information reliability and technical performance are the basis for shaping user value perception. For example, in AI-driven assistance systems, teachers are often more willing to tolerate a certain level of operational complexity as long as the system has substantial benefits in teaching (Gârdan et al., 2025). This result aligns with the research of Fan and Jiang (2024), who found similar patterns in their study of AI-assisted learning tools. Conversely, while technical support has a positive effect on perceived ease of use, it has no significant effect on perceived usefulness. This may be because teachers pay more attention to the chatbot’s professional capabilities and autonomous problem-solving characteristics rather than relying on external support services. This study has several limitations that should be acknowledged. First, the relatively small sample size may limit the representativeness and diversity of the findings, making it difficult to capture the full range of teachers’ perceptions and acceptance across different educational levels, subject backgrounds, or institutional settings. Second, the LLM-based chatbot used in this study is still in its early developmental stage, with system instability and incomplete functionality—such as interface issues and delayed responses—which may have interfered with participants’ experiences and affected their actual acceptance and usage behaviors. Third, as most participants were from economically developed regions, the sample may carry regional and cultural biases, reducing the extent to which the findings reflect the perspectives of educators in more diverse socioeconomic and cultural contexts. Fourth, although the study employed a mixed-methods approach incorporating both PLS-SEM and fsQCA, the qualitative interpretation derived from fsQCA configurations may lack depth due to the absence of complementary narrative data (e.g., interviews or open-ended responses). Without richer contextual insights from participants, the configurational findings—though informative—may not fully capture the underlying rationale or motivational reasoning behind teachers’ behavioral intentions. Building on the insights of this study, future research should consider several avenues for further exploration. First, expanding the sample to include a larger and more demographically diverse population, particularly teachers from underrepresented or rural regions, could provide a more nuanced and generalizable understanding of the factors influencing acceptance of AI ethics tools. Second, further technical development of the chatbot is essential. Future studies could assess the impact of specific chatbot improvements, such as adaptive learning features, multimodal interaction, or real-time feedback, on user satisfaction and learning outcomes. Third, cross-cultural or cross-regional comparative studies would be valuable in uncovering how contextual variables, such as cultural attitudes toward AI, policy environments, or institutional readiness, influence ethical AI adoption in education. Longitudinal designs could also offer insights into the sustained impact and evolving perceptions of AI-integrated teaching tools over time. Fourth, to enhance the explanatory power of fsQCA findings, future studies could incorporate qualitative interviews or case studies to triangulate results and provide deeper, context-rich interpretations of teachers’ acceptance patterns and ethical reasoning. Declarations Ethical Approval Statement This research protocol was approved by the Research Ethics Committee of the Graduate Institute, Wenzhou University (approval number: WZU-2025-0930C, approval date: September 30, 2025), and all procedures comply with the requirements of the Declaration of Helsinki. Informed consent statement All subjects signed a written informed consent form between October 15 and October 22, 2025, and voluntarily participated after fully understanding the purpose, procedures, risks and rights of the study. Data Availability The datasets generated during this study are fully available within the article and supplementary materials, which include the post-test questionnaire data on teachers' usage of the LLM-based chatbot. The supplementary files provide the raw dataset with basic coding applied. Author Contribution Chenchen Liu contributed to the Resources, Supervision, Funding acquisition, and Writing - Review & Editing. Yangying Yi contributed to the Conceptualization, Methodology, Formal analysis, Writing - Original Draft, and Writing - Review & Editing. Xinying Wang contributed to the Investigation, Data Curation, Writing - Original Draft, and Writing - Review & Editing. Haijie Wang contributed to the Conceptualization, Writing - Original Draft, and Writing - Review & Editing. Youmei Wang contributed to the Resources, Supervision, Project administration, and Writing - Review & Editing. References Al-Adwan, A. S., Li, N., Al-Adwan, A., Abbasi, G. A., Albelbisi, N. A., & Habibi, A. (2023). Extending the technology acceptance model (TAM) to Predict University Students’ intentions to use metaverse-based learning platforms. Education and Information Technologies, 28 (11), 15381-15413. https://doi.org/ 10.1007/s10639-023-11816-3 Al-Adwan, A. S., Nofal, M., Akram, H., Albelbisi, N. A., & Al-Okaily, M. (2022). Towards a Sustainable Adoption of E-Learning Systems: The Role of Self-Directed Learning. Journal of Information Technology Education: Research, 21 . https://doi.org/10.28945/4980 Al-Emran, M., Mezhuyev, V., & Kamaludin, A. (2018). Technology Acceptance Model in M-learning context: A systematic review. Computers & Education, 125 , 389-412. https:// doi.org/10.1016/j.compedu.2018.06.008 Aljuaid, H. (2024). The impact of artificial intelligence tools on academic writing instruction in higher education: A systematic review. Arab World English Journal (AWEJ) Special Issue on ChatGPT , 26-55. https://dx.doi.org/10.24093/ awej/ChatGPT.2 Almaiah, M. A., & Al Mulhem, A. (2019). Analysis of the essential factors affecting of intention to use of mobile learning applications: A comparison between universities adopters and non-adopters. Education and Information Technologies,24 (2),1433-1468.https://doi.org/10.1007/ s10639-018-9840-1 Almaiah, M. A., & Alamri, M. M. (2018). Proposing a new technical quality requirements for mobile learning applications. Journal of Theoretical and Applied Information Technology, 96 (19), 6955-6968. Almaiah, M. A., Jalil, M. A., & Man, M. (2016). Extending the TAM to examine the effects of quality features on mobile learning acceptance. Journal of Computers in Education, 3 , 453-485.https://doi.org/10.1007/s40692-016-0074-1 Almuhanna, M. A. (2024). Teachers’ perspectives of integrating AI-powered technologies in K-12 education for creating customized learning materials and resources. Education and Information Technologies , 1-29.https://doi.org/ 10.1007/s10639-024-13257-y Amhag, L., Hellström, L., & Stigmar, M. (2019). Teacher Educators’ Use of Digital Tools and Needs for Digital Competence in Higher Education. Journal of Digital Learning in Teacher Education, 35 (4), 203–220. https://doi.org/10.1080/21532974.2019.1646169 Arteaga Sánchez, R., Duarte Hueros, A., & García Ordaz, M. (2013). E‐learning and the University of Huelva: a study of WebCT and the technological acceptance model. Campus-Wide Information Systems, 30 (2), 135-160. https://doi.org/10.1108/ 10650741311306318 Ayanwale, M. A., Adelana, O. P., Molefi, R. R., Adeeko, O., & Ishola, A. M. (2024). Examining artificial intelligence literacy among pre-service teachers for future classrooms. Computers and education open, 6 (1)..https://doi.org/10.1016/j.caeo. 2024.100179 Baek, T. H., & Kim, M. (2023). Is chatgpt scary good? How user motivations affect creepiness and trust in generative artificial intelligence. Telematics and Informatics, 83 (3). https://doi.org/ 10.1016/j.tele.2023.102030 Bai, B., Wang, J., & Chai, C. S. (2021). Understanding Hong Kong primary school English teachers’ continuance intention to teach with ICT. Computer Assisted Language Learning, 34 (4), 528-551. https://doi.org/10.1080/09588221.2019. 1627459 Bakar-Corez, A., & Kocaman-Karoglu, A. (2024). E-dishonesty among postgraduate students and its relation to self-esteem. Education and Information Technologies, 29 (7), 8275-8300. https:// doi.org/10.1007/s10639-023-12105-9 Barrett, A. J., Pack, A., & Quaid, E. D. (2021). Understanding learners’ acceptance of high-immersion virtual reality systems: Insights from confirmatory and exploratory PLS-SEM analyses. Computers & Education, 169 (81). https:// doi.org/10.1016/j.compedu.2021.104214 Belda-Medina, J., & Calvo-Ferrer, J. R. (2022). Using Chatbots as AI Conversational Partners in Language Learning. Applied Sciences, 12 (17), 8427. https://doi.org/ 10.3390/app12178427 Bozkurt, A. (2023). Generative artificial intelligence (AI) powered conversational educational agents: The inevitable paradigm shift. Asian Journal of Distance Education, 18 (1), 198-204.https:// doi.org/10.5281/zenodo.7716416 Cheng, Y. M. (2012). Effects of quality antecedents on e‐learning acceptance. Internet Research, 22 (3), 361-390.https://doi.org/10.1108/10662241211235699 Chin, W. W. (1998). The partial least squares approach to structural equation modeling. Modern methods for business research, 295 (2), 295-336. Daneji, A. A., Ayub, A. F. M., & Khambari, M. N. M. (2019). The effects of perceived usefulness, confirmation and satisfaction on continuance intention in using massive open online course (MOOC). Knowledge Management & E-Learning, 11 (2), 201-214.https://doi.org/10.34105/ j.kmel.2019.11.010 Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management science, 35 (8), 982-1003.https://doi.org/ 10.1287/mnsc.35.8.982 DeLone, W. H., & McLean, E. R. (1992). Information systems success: The quest for the dependent variable. Information Systems Research, 38 (6), 373–384.https:// doi.org/10.1287/isre.3.1.60 DeLone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of information systems success: A 10-year update. Journal of Management Information Systems, 19 (4),9–30.https:// doi.org/10.1080/07421222.2003.11045748 Du, W., Liang, R. Y., & Liu, D. (2022). Factors influencing school teachers’ continuous usage intention of using VR technology for classroom teaching. Sage Open, 12 (3). https://doi.org/ 10.1177/21582440221114325 Dwivedi, Y. K., Wade, M. R., & Schneberger, S. L. (Eds.). (2011). Information systems theory: Explaining and predicting our digital society, vol. 1. https://doi.org/ 10.1007/978-1-4419-6108-2 Eke, D. O. (2023). Chatgpt and the rise of generative AI: Threat to academic integrity? Journal of Responsible Technology, 13 .https://doi.org/10.1016/j.jrt. 2023.100060 Farhan, W., Razmak, J., Demers, S., & Laflamme, S. (2019). E-learning systems versus instructional communication tools: Developing and testing a new e-learning user interface from the perspectives of teachers and students. Technology in Society, 59 . https://doi.org/10.1016/ j.techsoc.2019.101192. Fornell, C. & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research 18 , 39-50.https://doi.org/ 10.1177/002224378101800104 Gajic, N., & Boolaky, M. (2015). Impact of technical support on customer satisfaction: Case of automotive paints. SAGE Open, 5 (4).https://doi.org/10.1177/2158244 015611457 Gârdan, I. P., Manu, M. B., Gârdan, D. A., Negoiță, L. D. L., Paștiu, C. A., Ghiță, E., & Zaharia, A. (2025, February). Adopting AI in education: optimizing human resource management considering teacher perceptions. In Frontiers in Education, 10 . https://doi.org/10.3389/feduc. 2025.1488147 Ghavifekr, S., & Rosdy, W. A. W. (2015). Teaching and learning with technology: Effectiveness of ICT integration in schools. International journal of research in education and science, 1 (2), 175-191.https://doi.org/10.21890/ijres.23596 Giang, N. T. T., & Nga, P. T. H. (2024). The influence of information systems on student satisfaction: A study of perceived ease of use and perceived usefulness. Journal of Financial-Marketing Research, 15 (2), 62-72. https://doi.org/10.52932/jfm.vi2. 479 Gómez-Ramirez, I., Valencia-Arias, A., & Duque, L. (2019). Approach to M-learning acceptance among university students: An integrated model of TPB and TAM. International Review of research in open and distributed learning, 20 (3).https://doi.org/10.19173/irrodl.v20i4.4061 CopiedAn error has occu Grönroos, C. (1984). A service quality model and its marketing implications. European Journal of marketing, 18 (4), 36-44.https://doi.org/10.1108/EUM0000000004784 Hagendorff, T. (2020). The ethics of AI ethics: An evaluation of guidelines. Minds and machines, 30 (1), 99-120.https://doi.org/10.1007/s11023-020-09517-8 Hair, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2017). A primer on partial least squaresstructural equation modeling (PLS-SEM). International Journal of Research & Method in Education, 38 (2), 220–221.https://doi.org/10.1080/ 1743727X.2015.1005806 Han, J. H., & Sa, H. J. (2022). Acceptance of and satisfaction with online educational classes through the technology acceptance model (TAM): The COVID-19 situation in Korea. Asia Pacific Education Review, 23 (3), 403-415.https:// doi.org/10.1007/s12564-021-09716-7 Hsu, L. (2016). Examining EFL teachers’ technological pedagogical content knowledge and the adoption of mobile-assisted language learning: a partial least square approach. Computer Assisted Language Learning, 29 (8), 1287–1297. https://doi.org/10.1080/09588221.2016.1278024 Hu, H. H., Kandampully, J., & Juwaheer, T. D. (2009). Relationships and impacts of service quality, perceived value, customer satisfaction, and image: an empirical study. The service industries journal, 29 (2), 111-125.https://doi.org/10.1080/ 02642060802292932 Hu, Y. H. (2022). Effects and acceptance of precision education in an AI-supported smart learning environment. Education and Information Technologies,27 (2),2013-2037.https://doi.org/10.1007/ s10639-021-10664-3 Huang, Y. M. (2019). Examining students' continued use of desktop services: Perspectives from expectation-confirmation and social influence. Computers in Human Behavior, 96 , 23-31.https:// doi.org/10.1016/j.chb.2019.02.010 Ibrahim, A., & Shiring, E. (2022). The Relationship between Educators' Attitudes, Perceived Usefulness, and Perceived Ease of Use of Instructional and Web-Based Technologies: Implications from Technology Acceptance Model (TAM). International Journal of Technology in Education, 5 (4), 535-551.https:// doi.org/10.46328/ijte.285 Ipek, Z. H., Gozum, A. C., Papadakis, S., & Kallogianakis, M. (2023). Educational applications of the ChatGPT AI system: A systematic review research. Educational Process. International Journal, 12 (3), 26–55.https://doi.org/ 10.22521/edupij.2023.123.2 Jobin, A., Ienca, M., & Vayena, E. (2019). The global landscape of AI ethics guidelines. Nature machine intelligence, 1 (9), 389-399. https://doi.org/10.1038/s42256-019-0088-2 Kalankesh, L. R., Nasiry, Z., Fein, R. A., & Damanabi, S. (2020). Factors influencing user satisfaction with information systems: a systematic review. Galen Medical Journal, 9 . https:// doi.org/10.31661/gmj.v9i0.1686. Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., ... & Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and individual differences, 103 .https://doi.org/10.1016/j.lindif.2023.102274 Kazim, E., & Koshiyama, A. S. (2021). A high-level overview of AI ethics. Patterns, 2 (9). https://doi.org/10.1016/j.patter.2021.100314 Ke, C. H., Sun, H. M., & Yang, Y. C. (2012). Effects of User and System Characteristics on Perceived Usefulness and Perceived Ease of Use of the Web-Based Classroom Response System. Turkish Online Journal of Educational Technology, 11 (3), 128-143.https://files.eric.ed.gov/ fulltext/EJ989206.pdf Kim, N. J., & Kim, M. K. (2022, March). Teacher’s perceptions of using an artificial intelligence-based educational tool for scientific writing. In Frontiers in education, 7 . https://doi.org/10.3389/feduc.2022.755914 Kim, Y., & Lee, H. S. (2014). Quality, perceived usefulness, user satisfaction, and intention to use: An empirical study of ubiquitous personal robot service. Asian Social Science, 10 (11), 1.https://doi.org/10.5539/ass.v10n11p1 King, W., & He, J. (2006). A meta-analysis of the technology acceptance model. Information & Management, 43 , 740-755.https://doi.org/10.1016/j.im.2006. 05.003 Leppänen, P., George, G., & Alexy, O. (2023). When do novel business models lead to high performance? A configurational approach to value drivers, competitive strategy, and firm environment. Academy of management journal, 66 (1), 164-194.https://doi.org/10.5465/ amj.2020.0969 Lim, W. M., Gunasekara, A., Pallant, J. L., Pallant, J. I., & Pechenkina, E. (2023). Generative AI and the future of education: Ragnarök or reformation? A paradoxical perspective from management educators. The international journal of management education, 21 (2),.https://doi.org/10.1016/ j.ijme.2023.100790 Liu, I. F., Chen, M. C., Sun, Y. S., Wible, D., & Kuo, C. H. (2010). Extending the TAM model to explore the factors that affect intention to use an online learning community. Computers & education, 54 (2), 600-610. https://doi.org/10.1016/ j.compedu.2009.09.009 Ma, Q., & Liu, L. (2004). The technology acceptance model: A meta-analysis of empirical findings. Journal of Organizational and End User Computing, 16 (1), 59-72.https://doi.org/10.4018/ 978-1-59140-474-3.ch006 Machdar, N. M. (2016). The effect of information quality on perceived usefulness and perceived ease of use. Business and entrepreneurial review, 15 (2), 131-146. https://doi.org/10.25105/ ber.v15i2. 4630 Mahande, R. D., Jasruddin, J., & Nasir, N. (2019). Is success model for EDMODO e-learning user satisfaction through TAM on students. Journal of Educational Science and Technology, 5 (2), 140-152.https://doi.org/10.26858/est.v5i2.9575 Mouta, A., Torrecilla-Sánchez, E. M., & Pinto-Llorente, A. M. (2025). Comprehensive professional learning for teacher agency in addressing ethical challenges of AIED: Insights from educational design research. Education and Information Technologies, 30 (3), 3343-3387.https://doi.org/ 10.1007/s10639-024-12946-y Nelson, R. R., Todd, P. A., & Wixom, B. H. (2005). Antecedents of information and system quality: an empirical examination within the context of data warehousing. Journal of management information systems, 21 (4), 199-235. https://doi.org/ 10.1080/07421222.2005.11045823 Ngo, T. T. A., Tran, T. T., An, G. K., & Nguyen, P. T. (2024). ChatGPT for educational purposes: Investigating the impact of knowledge management factors on student satisfaction and continuous usage. IEEE Transactions on Learning Technologies, 17 , 1367–1378.https://doi.org/ 10.1109/TLT.2024.3383773 Nguyen, A., Ngo, H. N., Hong, Y., Dang, B., & Nguyen, B. P. T. (2023). Ethical principles for artificial intelligence in education. Education and information technologies, 28 (4), 4221-4241. Nikolopoulou, K. (2021). Mobile Devices and Mobile Learning in Greek Secondary Education: Policy, Empirical Findings and Implications. In Handbook for online learning contexts: Digital, mobile and open: Policy and practice (pp. 67-80). https://doi.org/10.1007/978-3-030-67349-9_6 Ohliati, J., & Abbas, B. S. (2019). Measuring students satisfaction in using learning management system. International Journal of Emerging Technologies in Learning (Online), 14 (4), 180-189. https://doi.org/10.3991/ijet.v14i04.9427 Onwuegbuzie, A. J., & Daniel, L. G. (2002). A framework for reporting and interpreting internal consistency reliability estimates. Measurement and evaluation in counseling and development, 35 (2), 89-103. https://doi.org/10.1080/07481756. 2002.12069052 Padayachee, I. (2017). Educator perceptions of virtual learning system quality characteristics. South African Computer Journal, 29 (3), 95-126. https://hdl. handle.net/10520/EJC-c5c8388c1 Pappas, I. O., & Woodside, A. G. (2021). Fuzzy-set Qualitative Comparative Analysis (fsQCA): Guidelines for research practice in Information Systems and marketing. International Journal of Information Management, 58 . https://doi.org/10.1016/ j.ijinfomgt.2021.102310 Patra, I., Alazemi, A., Al-Jamal, D., & Gheisari, A. (2022). The effectiveness of teachers’ written and verbal corrective feedback (CF) during formative assessment (FA) on male language learners’ academic anxiety (AA), academic performance (AP), and attitude toward learning (ATL). Language Testing in Asia, 12 (1), 19. https://doi.org/10.1186/s40468-022-00169-2 Rai, A., Lang, S. S., & Welker, R. B. (2002). Assessing the validity of IS success models: An empirical test and theoretical analysis. Information systems research, 13 (1), 50-69.https://doi.org/ 10.1287/isre.13.1.50.96 Rihoux, B., & Ragin, C. C. (2009). Configurational comparative methods: Qualitative comparative analysis (QCA) and related techniques. SAGE Publications, Inc .https://doi.org/10.4135/ 9781452226569 Scherer, R., Siddiq, F., & Teo, T. (2015). Becoming more specific: Measuring and modeling teachers' perceived usefulness of ICT in the context of teaching and learning. Computers & Education, 88 , 202-214. https://doi.org/10.1016/j. compedu.2015.05.005 Scherer, R., Siddiq, F., & Tondeur, J. (2019). The technology acceptance model (TAM): A meta-analytic structural equation modeling approach to explaining teachers’ adoption of digital technology in education. Computers & education, 128 , 13-35.https://doi.org/10.1016/ j.compedu.2018.09.009 Seddon, P. B. (1997). A respecification and extension of the DeLone and McLean model of IS success. Information systems research, 8 (3), 240-253. https://doi.org/ 10.1287/isre.8.3.240 Shen, X., & Liu, J. (2022). Analysis of factors affecting user willingness to use virtual online education platforms. International Journal of Emerging Technologies in Learning, 17 (1), 74-89. https://doi.org/10.3991/ijet.v17i01.28713 Siafis, V., Rangoussi, M., & Psaromiligkos, Y. (2024). Recommender Systems for Teachers: A Systematic Literature Review of Recent (2011–2023) Research. Education Sciences, 14 (7), 723.https://doi.org/10.3390/educsci14070723 Teo, T. (2010). A path analysis of pre-service teachers' attitudes to computer use: applying and extending the technology acceptance model in an educational context. Interactive Learning Environments, 18 (1), 65-79. https://doi.org/10. 1080/10494820802231327 Uddin, M.M. (2024). Rejection or integration of AI in academia: determining the best choice through the Opportunity Cost theoretical formula. Discover Education , 3 (249).https://doi.org/10.1007/ s44217-024-00349-7 Waheed, M., & Jam, F. A. (2010). Teacher’s intention to accept online education: Extended TAM model. Interdisciplinary Journal of Contemporary Research in Business, 2 (5), 330-344. Wang, S., Sun, Z., & Chen, Y. (2023). Effects of higher education institutes’ artificial intelligence capability on students’ self-efficacy, creativity and learning performance. Education and Information Technologies, 28 (5), 4919–4939. https://doi.org/10.1007/s10639-022-11338-4. Wang, Y., Liu, C., & Tu, Y. F. (2021). Factors affecting the adoption of AI-based applications in higher education. Educational Technology & Society, 24 (3), 116-129.https://www.jstor.org/ stable/27032860 Wiese, L. J., Patil, I., Schiff, D. S., & Magana, A. J. (2025). AI Ethics Education: A Systematic Literature Review. Computers and Education: Artificial Intelligence, 8 . https://doi.org/10.1016/ j.caeai.2025.100405 Xu, J., Benbasat, I., & Cenfetelli, R. T. (2013). Integrating service quality with system and information quality: An empirical test in the e-service context. MIS quarterly, 37 (3), 777-794. https://www.jstor.org/stable/43825999 Zhang, C., Schießl, J., Plößl, L., Hofmann, F., & Gläser-Zikuda, M. (2023). Acceptance of artificial intelligence among pre-service teachers: a multigroup analysis. International Journal of Educational Technology in Higher Education, 20 (1), 49.https://doi.org/10.1186/s41239- 023-00420-7 Zheng, Y., Zhao, K., & Stylianou, A. (2013). The impacts of information quality and system quality on users' continuance intention in information-exchange virtual communities: An empirical investigation. Decision support systems, 56 , 513-524.https://doi.org/10.1016/j.dss.2012.11.008 Additional Declarations No competing interests reported. Supplementary Files RawDataset.xlsx Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 07 May, 2026 Reviews received at journal 02 May, 2026 Reviewers agreed at journal 25 Apr, 2026 Reviewers agreed at journal 09 Apr, 2026 Reviews received at journal 05 Apr, 2026 Reviewers agreed at journal 27 Mar, 2026 Reviewers invited by journal 16 Mar, 2026 Editor assigned by journal 16 Mar, 2026 Editor invited by journal 18 Feb, 2026 Submission checks completed at journal 15 Feb, 2026 First submitted to journal 15 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8618094","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":606926308,"identity":"cbdf6714-4032-4cc0-86d1-637ae255c7cd","order_by":0,"name":"Chenchen Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYBACPgYG9g8SFWz1jO3NBw58qCBCCxsIWZzhS2DuOZZ4cMYZYrVUtsklsM/IMT7M20KMFvbDzx7cYDPL4+058+EAbwODPL/YAQJaeNLMDWfwpBVLtvduOCC5g8Fw5uwEAlokeBikJSSOMW7sObvhgOEZhgSD28Ro+WPwn3H/jZwHBxLbiNPCJiGRwJbYOCOH4cBBorTwpBkbSBxgM2bsOWZwsOGMBGG/8LMffvhA8h+bHDAqH3/+U2Ejzy9NQAs6kCBN+SgYBaNgFIwC7AAARqVHEiclfB4AAAAASUVORK5CYII=","orcid":"","institution":"Wenzhou University","correspondingAuthor":true,"prefix":"","firstName":"Chenchen","middleName":"","lastName":"Liu","suffix":""},{"id":606926309,"identity":"be13559f-a0f6-4705-b7aa-f106ad47fd9c","order_by":1,"name":"Yangying Yi","email":"","orcid":"","institution":"Wenzhou University","correspondingAuthor":false,"prefix":"","firstName":"Yangying","middleName":"","lastName":"Yi","suffix":""},{"id":606926310,"identity":"a6753ac4-7968-466a-8b6b-95a56f6ad433","order_by":2,"name":"Xinying Wang","email":"","orcid":"","institution":"Wenzhou University","correspondingAuthor":false,"prefix":"","firstName":"Xinying","middleName":"","lastName":"Wang","suffix":""},{"id":606926311,"identity":"5ab2ca6a-6365-4d6a-a059-94a9e3576523","order_by":3,"name":"Haijie Wang","email":"","orcid":"","institution":"Central China Normal University","correspondingAuthor":false,"prefix":"","firstName":"Haijie","middleName":"","lastName":"Wang","suffix":""},{"id":606926312,"identity":"87cbc8fd-4be0-41f4-b970-855fe4752530","order_by":4,"name":"Youmei Wang","email":"","orcid":"","institution":"Wenzhou University","correspondingAuthor":false,"prefix":"","firstName":"Youmei","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2026-01-16 11:25:37","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8618094/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8618094/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105763812,"identity":"c289fe24-10ce-46e5-becf-1d329e0125ac","added_by":"auto","created_at":"2026-03-30 19:18:32","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":139032,"visible":true,"origin":"","legend":"\u003cp\u003eTheoretical confirmatory model\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8618094/v1/cb3cc897cb083be26e9ce453.png"},{"id":105763814,"identity":"b414ed89-2ea0-4453-8883-d4ba47e50dea","added_by":"auto","created_at":"2026-03-30 19:18:32","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":542243,"visible":true,"origin":"","legend":"\u003cp\u003eLearning interface of the chatbot.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8618094/v1/49ad8053eb5e75dc89d45f3c.png"},{"id":105763815,"identity":"102abe6a-ad8a-48ad-8378-65e66395b0ad","added_by":"auto","created_at":"2026-03-30 19:18:32","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":417203,"visible":true,"origin":"","legend":"\u003cp\u003eThe modules and functions of the chatbot\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8618094/v1/b8c719f3be48384c8bfd1905.png"},{"id":106401523,"identity":"909dae02-76c2-4a32-ac4d-4728c5241485","added_by":"auto","created_at":"2026-04-08 09:06:22","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":277538,"visible":true,"origin":"","legend":"\u003cp\u003eConfirmatory model results\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8618094/v1/24303ddd8b385bdfd90f503d.png"},{"id":106414874,"identity":"1c1eb7cd-cc67-4b71-9ede-26214aa1b82b","added_by":"auto","created_at":"2026-04-08 10:29:31","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2452265,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8618094/v1/64714931-9134-40bf-8ae6-5efc5d06acc9.pdf"},{"id":106092936,"identity":"9a36cbaa-815a-4dde-acbd-ff68716691c2","added_by":"auto","created_at":"2026-04-03 11:31:04","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":27484,"visible":true,"origin":"","legend":"","description":"","filename":"RawDataset.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8618094/v1/abeebf6e9f27760d7a5d0b43.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Understanding AI ethics in education with a LLM-based chatbot: Evidence from PLS-SEM and fsQCA analyses","fulltext":[{"header":"Introduction","content":"\u003cp\u003eWith the continuous innovation of technology, Generative Artificial Intelligence (GenAI) is becoming increasingly and widely used in the field of education, and its rapid integration in educational environments is profoundly changing the way of teaching and learning. While GenAI offers significant opportunities for enhancing teaching efficiency and personalized instruction (Bozkurt, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), it also raises a range of key ethical questions about transparency, fairness, accountability, and changing roles of educators (Jobin et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Kazim \u0026amp; Koshiyama, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Nguyen et al., \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Research suggests that over-reliance on AI can impair the ability of teachers and students to think independently and creativity, leading to academic misconduct such as cheating, plagiarism, or using AI to complete assignments and write essays (Bakar-Corez \u0026amp; Kocaman-Karoglu, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Kasneci et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Additionally, AI-generated answers and content can hinder critical thinking for teachers and students, bypassing traditional learning processes, thereby damaging their academic development and the credibility of the education system (Ipek et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Additionally, the information generated by GenAI tools may be inaccurate or biased, further increasing the risk of academic dishonesty and copyright infringement (Baek \u0026amp; Kim, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Eke, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In response to the growing ethical issue of AI, studies have pointed out that a complete rejection of AI is not a viable solution (Lim et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Many scholars and educational institutions advocate for an open and cautious approach, emphasizing the development of clear frameworks and clear guiding principles to harness the full potential of AI while effectively mitigating its potential risks (Uddin, \u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e While numerous ethical guidelines and policies for AI have emerged in recent years, these norms often remain abstract and lack concrete implementation guidance. Thus, teachers struggle to clearly identify and respond to the ethical risks and responsibilities of using AI in educational contexts (Nikolopoulou, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). At the same time, teachers play a vital role in guiding students to use AI responsibly and consciously, their ethical awareness and literacy are essential for the sustainable integration of AI in education (Ayanwale et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). However, the opportunities for teachers to participate in AI ethics learning are still limited and somehow ineffective. The lack of structured professional development pathways limits educators\u0026rsquo; ability to meaningfully engage in and contribute to discussions and practices related to AI ethics (Lim et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This suggests that existing policies and guidelines alone are insufficient to address current challenges, highlighting the need for targeted training mechanisms to systematically enhance teachers\u0026rsquo; cognition, understanding, and practical competence in AI ethics (Mouta et al., \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Conventional training approaches frequently prove inadequate in addressing the specific demands of this domain. On the one hand, conventional training is frequently detached from practical teaching contexts and lacks situational relevance. On the other hand, its content often lacks flexibility and continuity, limiting its effectiveness in keeping pace with fast-evolving technological developments.(Hagendorff, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wiese et al., \u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Accordingly, it is imperative to design innovative and scalable training models that facilitate teachers\u0026rsquo; continuous professional development and promote deep, context-aware ethical reflection. To address this need, this study introduces chatbot, a large language model-based chatbot developed via the LangChain framework, designed to support teachers in enhancing their ethical literacy in the context of artificial intelligence.\u003c/p\u003e \u003cp\u003eTo understand teachers\u0026rsquo; intention to use the chatbot and the factors influencing it, this study adopts the Technology Acceptance Model (TAM) to examine how key variables\u0026mdash;such as perceived usefulness and perceived ease of use\u0026mdash;affect their usage intentions. TAM has been widely substantiated in many fields to effectively explain user behavior in adopting information systems, particularly focusing on perceived usefulness and perceived ease of use as key determinants factors affecting user intention and actual usage behavior (Barrett et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Davis, 1989; Wang et al., \u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). But as research advances further, more and more scholars have begun to pay attention to other potential influencing factors other than the TAM model (Barrett et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Kim \u0026amp; Lee, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Studies have shown that system quality is also an important factor affecting user adoption behavioral intention to adopt technology (Shen \u0026amp; Liu, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Xu et al., \u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Zheng et al., \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), high-quality systems often enhance user trust, satisfaction, and overall user experience, thereby increasing their willingness to adopt and continue using them (Almaiah \u0026amp; Mulhem, 2019; Al-Adwan et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This study introduces a specific dimension of AI ethics education based on TAM, focusing on three key factors affecting the system quality of the chatbot: content accuracy, system functionality and stability, and technical support. These factors are hypothesized to affect users\u0026rsquo; perceptions of usefulness and ease of use, further affecting their behavioral intentions to continue using the chatbot. By integrating PLS-SEM and fsQCA, this study comprehensively reveals the causal relationships and multiple configuration paths that affect teachers\u0026rsquo; willingness to use AI ethics learning tools, providing a new theoretical perspective for understanding their motivations.\u003c/p\u003e"},{"header":"Literature review","content":"\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTechnology acceptance model\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Technology Acceptance Model (TAM) was first introduced by Davis (1989) to explore the degree of user acceptance of emerging technologies. As a maturing theoretical model, TAM has been widely used to understand the key factors influencing user adoption and sustained use of technological innovations (Al-Emran et al., 2018; Teo, 2010). Over the past decades, it has been successfully applied in various fields, including mobile learning (Almaiah et al., 2016; G\u0026oacute;mez-Ramirez et al., 2019), smart learning environments (Hu, 2022), virtual learning platforms (Al-Adwan et al., 2023), and online education systems (Han \u0026amp; Sa, 2022; Waheed \u0026amp; Jam, 2010). With the expansion of its application contexts, an increasing number of studies have explored the \u0026nbsp;applicability of the TAM model in the context of diversification, and proposed extensions to its original structure. The model emphasizes two key constructs:\u0026nbsp;Perceived usefulness (PU) and perceived ease of use (PEU), which are considered crucial in determining users\u0026rsquo;\u0026nbsp;attitudes and behavioral intentions toward a particular technology. Perceived usefulness refers to the degree to which users believe that a technology can improve their performance, while perceived ease of use refers to the degree to which users find the technology free from effort (Davis, 1989; Liu et al., 2010). These perceptions directly influence users\u0026rsquo; intention to use the technology, which serves as the primary outcome variable in TAM. It is widely regarded as a strong predictor of actual usage behavior and is critical in assessing the effectiveness of educational technology adoption (Barrett et al., 2021; Scherer et al., 2019; Zhang et al., 2023).\u003c/p\u003e\n\u003cp\u003eNumerous studies have consistently shown that among the key constructs of the TAM, PU is typically a strong predictor of teachers\u0026rsquo; intentions to adopt new technologies. Meta-analyses by Ma and Liu (2004) and King and He (2006) show that PU has a wide range of coefficient values, indicating its strong impact on users\u0026rsquo; intentions to use various technological tools. This effect is particularly evident in the educational context, where PU serves as a crucial determinant of teachers\u0026rsquo; intentions to adopt new technologies. When teachers perceive that a technology can enhance instructional effectiveness, streamline teaching processes, or facilitate student learning, they are more likely to adopt it in their teaching practices (Bai et al., 2021; Scherer et al., 2015). Similarly, complementing the role of PU, PEU is likewise supported by educational technology research as a key predictor of teachers\u0026rsquo; ITU. For instance, Teo (2010) found that PEU had a significant positive influence on pre-service teachers\u0026rsquo; intention to use educational software, indicating that ease of system interaction is an important motivational factor. Weng et al. (2018) also confirmed in their TAM-based study on elementary school teachers\u0026rsquo; use of multimedia materials that PEU not only has a direct effect on ITU, but also exerts an indirect effect through attitude toward use. These findings provide the basis for the following two hypotheses of this study：\u003c/p\u003e\n\u003cp\u003eH1. PEU will have a positive and significant effect on teachers\u0026rsquo; ITU of the chatbot.\u003c/p\u003e\n\u003cp\u003eH2. PU will have a positive and significant effect on teachers\u0026rsquo; ITU of the chatbot.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eThe quality of learning system\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Information System Success Model (IS Success Model), proposed by DeLone and McLean (1992), is one of the most widely adopted frameworks for evaluating the effectiveness of information systems. This model conceptualizes system success through three primary dimensions: information quality, system quality, and service quality. Subsequently, Rai et al. (2002) provided empirical validation demonstrating that the three dimensions of the information system success model are key constructs and factors influencing IS usage and performance. This study modifies the original IS success model to align more closely with the educational application scenario of the AI ethics learning tool. Specifically, information quality is revised as \u0026ldquo;content accuracy\u0026rdquo;, referring to the relevance and correctness of the ethical content generated by the chatbot; system quality is reformulated as \u0026ldquo;system functionality and stability\u0026rdquo;, capturing the chatbot\u0026rsquo;s operational reliability and technical performance; and service quality is renamed \u0026ldquo;technical support\u0026rdquo;, assessing the quality and responsiveness of assistance provided to users. This study draws on Cheng\u0026rsquo;s (2012) empirical findings that information quality, system quality, and service quality are the key quality antecedent variables influencing users\u0026rsquo; acceptance of e-learning systems. Based on the research conclusions of Cheng (2012), this study defines these three quality dimensions as external antecedents affecting teachers\u0026rsquo; PU and PEU. In the proposed research model, these quality dimensions function as external variables that influence teachers\u0026rsquo; system perceptions and ultimately shape their intention to adopt the AI-based ethical learning tool.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eContent Accuracy\u003cstrong\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eContent accuracy refers to the ability of a system to provide correct, trustworthy, and up-to-date information (DeLone \u0026amp; McLean, 2003; Sassano et al., 2010). As a key component of information quality, content accuracy is fundamental in shaping users\u0026rsquo; trust and overall evaluation of a technology system, particularly in knowledge-intensive environments (Nelson et al., 2005). In AI-supported tools, maintaining high content accuracy is particularly important to ensure teaching reliability and professional credibility as content is dynamically generated (Almuhanna, 2024). Seddon (1997) points out that the user\u0026rsquo;s perception of the usefulness of a system is greatly influenced by the quality of the information it provides. Machdar (2016) further emphasized that when users rely on system-generated content for decision-making, higher information quality makes the system more likely to be perceived as useful and trustworthy. For teachers, the content accuracy of AI-generated tools significantly enhances their confidence in applying these tools to teaching tasks, such as writing assistants or instructional simulations (Aljuaid, 2024; Kim \u0026amp; Kim, 2022). In addition, studies have found that teachers are more inclined to adopt systems that provide accurate and highly relevant resources in AI-based intelligent recommendation systems (Siafis et al., 2024).\u003c/p\u003e\n\u003cp\u003eBeyond fostering users\u0026rsquo; perceived usefulness, content accuracy also facilitates perceived ease of use. When information is accurate and clearly presented, users can more easily understand how to operate the system, thereby enhancing their perception of its ease of use (Almaiah \u0026amp; Al Mulhem, 2019). For example, in digital assessment systems, when teachers receive accurate and well-structured feedback, they are less likely to need additional explanation or manual correction, which reduces disruptions during use and enhances the system\u0026rsquo;s operability (Patra et al., 2022). Moreover, in language education, teachers generally believe that high content accuracy as a key factor in reducing the need for repeated verification of system outputs. This reduction in operational pressure and technical anxiety further strengthens their perception of system ease of use (Belda-Medina \u0026amp; Calvo-Ferrer, 2022; Hsu, 2016).\u003c/p\u003e\n\u003cp\u003eBased on these insights, the following hypotheses are proposed:\u003c/p\u003e\n\u003cp\u003eH3. CA of the chatbot will be a positive and significant predictor of PU.\u003c/p\u003e\n\u003cp\u003eH4. CA of the chatbot will be a positive and significant predictor of PEU.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSystem functionality and stability \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSystem functionality and stability refer to the overall technical performance demonstrated by the system during its operation, including operational stability, functional reliability, system response speed, interface usability, and cross-platform compatibility (Almaiah \u0026amp; Alamri, 2018; DeLone \u0026amp; McLean, 2003; Giang \u0026amp; Nga, 2024; Kim \u0026amp; Lee, 2014; Ohliati \u0026amp; Abbas, 2019). This variable examines the system\u0026rsquo;s ability to provide stable, uninterrupted service, ensure smooth access to its functions, and support users in completing tasks efficiently. System functionality and stability are key characteristics of a information system that can significantly influence users\u0026rsquo; attitudes and behaviors towards its use (Almaiah \u0026amp; Al Mulhem, 2019).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA system with comprehensive features and high stability can provide a more efficient and smoother user experience, fostering users\u0026rsquo; belief that it can effectively meet their needs and thereby enhancing its perceived usefulness (Kim \u0026amp; Lee, 2014). According to Farhan et al. (2019), system quality dimensions\u0026mdash;including reliability, response time, and interface design\u0026mdash;have a significant impact on teachers\u0026rsquo;\u0026nbsp;perceived usefulness of school e-learning systems.\u0026nbsp;When a system is highly functional and stable, teachers are more likely to trust the technology and believe in its potential to improve teaching organization, classroom interaction, and student engagement. As Du et al. (2022) point out,\u0026nbsp;when evaluating the usefulness of virtual reality technology in classroom teaching, teachers pay particular attention to whether the system can maintain continuous and stable operation and whether it can support diverse teaching situations.\u003c/p\u003e\n\u003cp\u003eSystem functionality and stability play a critical role in shaping users\u0026rsquo; PEU. A successful online learning system should be evaluated in terms of user-friendliness and the effectiveness of feedback mechanisms to ensure that learners can operate smoothly and receive timely support (Giang \u0026amp; Nga, 2024; Mahande et al., 2019). For teachers, A reliable and user-friendly system allows them to navigate it easily, build teaching confidence, and minimize frustration (Ghavifekr \u0026amp; Rosdy, 2015; Padayachee, 2017). Building on this, Ke et al. (2012) demonstrated that systems with well-designed interfaces and stable performance can help teachers master and use the technology to teach more efficiently and increase student participation in the classroom.\u003c/p\u003e\n\u003cp\u003eBased on these insights, the following hypotheses are proposed:\u003c/p\u003e\n\u003cp\u003eH5. SFS of the chatbot will positively and significantly predict PU.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eH6. SFS of the chatbot will positively and significantly predict PEU.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eTechnical Support\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTechnical support refers to the level of service provided by the information system to assist users. It includes the timeliness of responses, the effectiveness of problem resolution, the professionalism of support staff, and the adequacy of user training and guidance (Dwivedi et al., 2011; Gr\u0026ouml;nroos, 1984; Ting et al., 2011). During system usage, users may encounter system errors, difficulties in operating specific functions, or incompatibility due to system updates. Timely intervention by technical support can significantly reduce barriers to use and enhance user trust and satisfaction with the system (Gajic \u0026amp; Boolaky, 2015; Hu et al., 2009; Kalankesh et al., 2020).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTechnical support has a positive impact on PU by improving the reliability and operation guarantee of the system and enhancing the use\u0026rsquo;s trust in the actual utility of the system. In educational contexts, teachers are often not technical experts. When faced with unfamiliar teaching tools, timely and professional technical assistance can help them better perceive the system\u0026rsquo;s support for their teaching objectives (Ghavifekr \u0026amp; Rosdy, 2015; Padayachee, 2017). In addition, empirical studies on teachers\u0026rsquo; use of virtual reality technology in the classroom have shown that technical support has a significant impact on the perceived usefulness of their systems, which in turn further affects teachers\u0026rsquo; user satisfaction (Du et al., 2022). This further shows that technical support, as a core element of service quality, is a key factor in shaping the perception of teachers\u0026rsquo; system usefulness.\u003c/p\u003e\n\u003cp\u003eAt the same time, technical support also plays an important role in the user\u0026rsquo;s PEU. Timely and effective assistance can reduce cognitive and technical barriers during operation, allowing teachers to use system functions more smoothly (Arteaga S\u0026aacute;nchez et al., 2013; Ibrahim \u0026amp; Shiring, 2022).\u0026nbsp;Empirical research further confirms this view. In a study on the use of WebCT teaching systems by college teachers, Kim and Lee (2014) found that the continuous technical support of the\u0026nbsp;chatbot\u0026nbsp;significantly reduced teachers\u0026apos; uncertainty in the face of technology, thereby improving their overall perception of the ease of use of the system. In addition, Dwivedi et al. (2011) pointed out that technical support improves system usability by providing assistance in key usage links, thereby enhancing the user\u0026rsquo;s perception of system learnability and intuitiveness.\u003c/p\u003e\n\u003cp\u003eBased on these insights, the following hypotheses are proposed:\u003c/p\u003e\n\u003cp\u003eH7.TS of the chatbot will be a positive and significant predictor of PU.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eH8.TS of the chatbot will be a positive and significant predictor of PEU.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTheoretical confirmatory model\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBased on the synthesis of prior studies outlined in Subsections 2.1 and 2.2, this study formulates a confirmatory theoretical framework (see Figure 1) to empirically explore teachers\u0026rsquo; acceptance of the chatbot. In the proposed model, content accuracy, system functionality and stability, and technical support are included as external variables that influence teachers\u0026rsquo; perceived ease of use and perceived usefulness, which in turn affect their intention to use the chatbot. This model provides a structured framework for understanding key factors that drive technology adoption.\u003c/p\u003e"},{"header":"Research methods","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eSample and data collection\u003c/h2\u003e \u003cp\u003eThe research employed a stratified random sampling method, selecting, teachers from a public university located in southeastern China as the research subjects. A questionnaire was formulated after referring to previous studies and consulting with experts to ensure the validity of the questionnaire. As depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, the three structures are named \u0026ldquo;Content accuracy,\u0026rdquo; \u0026ldquo;System function and stability\u0026rdquo;, and \u0026ldquo;Technical support.\u0026rdquo; All items were evaluated using a five-point Likert scale, with a score range from 1 (strongly disagree) to 5 (strongly agree).\u003c/p\u003e \u003cp\u003eTo ensure the reliability and validity of the collected data, the research team set a critical criterion: all teachers who participated in the survey must have had at least one month of experience using the chatbot. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows a participant interacting with the chatbot. This criterion not only helps in obtaining feedback based on actual usage experiences but also reduces cognitive biases due to short usage times, ensuring that the research results accurately reflect teachers\u0026rsquo; technical acceptance and usage intentions of the chatbot. Under the strict adherence to the above sampling principles and standards, a total of 168 questionnaires were initially collected in this study. After a meticulous screening process, including the elimination of questionnaires with blank filling, missing key information, and arbitrary filling, 142 valid questionnaires were finally retained for further analysis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eChatbot\u003c/h2\u003e \u003cp\u003eThis study presents the design of an AI-assisted chatbot for ethical learning and reflection, which integrates a large language model (LLM) through API access. The chatbot leverages the LLM\u0026rsquo;s natural language processing capabilities to facilitate structured dialogue and critical engagement with ethical issues in educational contexts, aiming to support the development of teachers\u0026rsquo; ethical awareness and reasoning.\u003c/p\u003e \u003cp\u003eThe chatbot is designed around three main functions: knowledge expansion, case study, and post-learning evaluation to support the acquisition, application, and evaluation of ethical knowledge. Through the knowledge expansion module, teachers can access standardized explanations of fundamental concepts, principles, and common issues related to AI ethics by interacting with a large language model, thereby enhancing their theoretical understanding. The case-based learning module offers a rich repository of educational ethics cases, enabling teachers to engage with concrete scenarios, analyze ethical dilemmas, and internalize ethical knowledge by applying theory to practice. The post-learning assessment module dynamically generates evaluation items to assess teachers\u0026rsquo; mastery of ethical knowledge, providing personalized feedback that fosters ongoing reflection and professional growth. Below is a detailed explanation of the functions of each module (see Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eData analysis\u003c/h2\u003e \u003cp\u003eThis study employed SmartPLS 4.0 for Partial Least Squares Structural Equation Modeling (PLS-SEM), complemented by fuzzy-set Qualitative Comparative Analysis (fsQCA), to comprehensively analyze the factors influencing teachers\u0026rsquo; intention to continue using the chatbot.\u003c/p\u003e \u003cp\u003ePLS-SEM was selected for its robust handling of small sample sizes (n\u0026thinsp;=\u0026thinsp;142) and its ability to estimate complex structural models with multiple latent constructs, such as content accuracy, system functionality and stability, technical support, perceived usefulness, perceived ease of use, and intention to use. This method is suitable for exploratory research and theoretical development with low distribution assumptions (Hair et al., 2021).\u003c/p\u003e \u003cp\u003eTo enrich the findings and address the complexity of configurations in the user decision-making process, the fsQCA method is also introduced. PLS-SEM is primarily used to test linear net effects between constructs, while fsQCA is able to identify multiple causal paths that lead to the outcome variable\u0026mdash;intention to use\u0026mdash;through a combination of conditions such as high content accuracy, functional stability, and ease of use. This combined strategy provides both variance-based and set theory-based insights. Data analysis is divided into three steps:\u003c/p\u003e \u003cp\u003e(1) Measurement model assessment, involving the evaluation of indicator loadings, composite reliability, and average variance extracted to ensure the validity and reliability of latent constructs.\u003c/p\u003e \u003cp\u003e(2) Structural model assessment, including estimation of path coefficients and explanatory power (R\u0026sup2;) to test hypothesized relationships among constructs.\u003c/p\u003e \u003cp\u003e(3) FsQCA, which was used to identify sufficient configurations of causal conditions that lead to high continuance intention, thus capturing the multifaceted and equifinal nature of user acceptance in ethical AI learning tools.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eMeasurement model estimation\u003c/h2\u003e \u003cp\u003eThe reliability and validity of the latent variables were tested based on previous studies evaluating the measurement model (Hair et al., \u003cspan class=\"CitationRef\"\u003e2017\u003c/span\u003e), as shown in Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. All individual projects have a reliability (load) of 0.635 or higher, meeting the criteria for reliability of individual projects (Hajiar, 2018). Internal consistency was assessed using both composite reliability (CR) and Cronbach’s alpha. In this study, the Cronbach’s alpha ranged from 0.833 to 0.914, exceeding the commonly accepted threshold of 0.7, thereby indicating satisfactory reliability (Onwuegbuzie \u0026amp; Daniel, \u003cspan class=\"CitationRef\"\u003e2002\u003c/span\u003e). Additionally, the CR values for all constructs fell between 0.885 and 0.933, which is well above the minimum recommended level of 0.70 and within the ideal range of 0.70 to 0.90, reflecting strong internal consistency. Convergent validity (CV) was examined through the Average Variance Extracted (AVE). The AVE scores in this study varied between 0.523 and 0.768, all surpassing the benchmark value of 0.50 as proposed by Chin (\u003cspan class=\"CitationRef\"\u003e1998\u003c/span\u003e) and Fornell and Larcker (\u003cspan class=\"CitationRef\"\u003e1981\u003c/span\u003e). These results suggest that the measurement model demonstrates adequate convergent validity.\u003c/p\u003e \u003cp\u003eDiscriminative validity refers to the ability of a variable to be distinguishable from other variables to some extent. The Fornell-Larcker criterion is a method used to test discriminative validity. According to this criterion, the condition for determining validity is that the square root of the mean variance extraction for each variable should be greater than the correlation coefficient between that variable and any other variable (Fornell \u0026amp; Larcker, \u003cspan class=\"CitationRef\"\u003e1981\u003c/span\u003e). The results in Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e show that the square root of the average variance extracted for each variable is greater than its highest correlation coefficient with other variables, thus meeting the requirements for discriminative validity.\u003c/p\u003e \u003cp\u003eIn conclusion, this study systematically tests the reliability and validity of the measurement model, and the results show that the factor load of each variable exceeds 0.7, and the internal consistency and aggregate validity meet the standard. This indicates that the measurement model has high reliability and validity, and is suitable for subsequent structural model analysis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003ctable id=\"Tab1\" border=\"1\"\u003e \u003ccaption\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCFA results for reliability and validity of measurement items\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003c/colgroup\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\"\u003e \u003cp\u003eConstruct\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eOriginal\u003c/p\u003e \u003cp\u003eitems\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eFactor\u003c/p\u003e \u003cp\u003eloadings\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eCronbach’s\u003c/p\u003e \u003cp\u003ealpha\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eComposite\u003c/p\u003e \u003cp\u003ereliability\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eAverage variance extracted (AVE)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" rowspan=\"5\"\u003e \u003cp\u003eContent accuracy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.743\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"5\"\u003e \u003cp\u003e0.871\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"5\"\u003e \u003cp\u003e0.907\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"5\"\u003e \u003cp\u003e0.66\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0 842\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.781\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.859\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" rowspan=\"7\"\u003e \u003cp\u003eSystem functionality and stability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFA1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.716\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"7\"\u003e \u003cp\u003e0.848\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"7\"\u003e \u003cp\u003e0.885\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"7\"\u003e \u003cp\u003e0.523\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFA2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.725\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFA3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.752\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFA4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.751\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFA5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.721\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFA6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFA7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.727\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" rowspan=\"4\"\u003e \u003cp\u003eTechnical support\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTS1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.758\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"4\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"4\"\u003e \u003cp\u003e0.889\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"4\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTS2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.827\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTS3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.852\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTS4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.827\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" rowspan=\"7\"\u003e \u003cp\u003ePerceived ease of use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eEU1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0 635\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"7\"\u003e \u003cp\u003e0.892\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"7\"\u003e \u003cp\u003e0.916\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"7\"\u003e \u003cp\u003e0.612\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eEU2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.832\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eEU3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.836\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eEU4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.829\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eEU5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.834\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eEU6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.788\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eEU7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.697\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" rowspan=\"6\"\u003e \u003cp\u003ePerceived usefulness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eU1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.847\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"6\"\u003e \u003cp\u003e0.914\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"6\"\u003e \u003cp\u003e0.933\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"6\"\u003e \u003cp\u003e0.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eU2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.803\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eU3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.815\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eU4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.878\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eU5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.789\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eU6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.885\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" rowspan=\"4\"\u003e \u003cp\u003eIntention to use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eWCU1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.865\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"4\"\u003e \u003cp\u003e0.899\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"4\"\u003e \u003cp\u003e0.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"4\"\u003e \u003cp\u003e0.768\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eWCU2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.891\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eWCU3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.875\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eWCU4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.875\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003ctable id=\"Tab2\" border=\"1\"\u003e \u003ccaption\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDiscriminant validity results\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003c/colgroup\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eContent accuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003ePerceived ease of use\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eSystem functionality and stability\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eTechnical support\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003ePerceived usefulness\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eIntention to use\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eContent accuracy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.813\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePerceived ease of use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.633\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.782\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSystem functionality and stability\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.644\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.723\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTechnical support\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.404\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.497\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.509\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.817\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePerceived usefulness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.606\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.871\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.661\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.457\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.837\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eIntention to use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.653\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.648\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.397\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.716\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.876\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eStructural model estimation\u003c/h2\u003e \u003cp\u003ePLS-SEM analyzes a structural model by the indicators of: (1)R\u003csup\u003e2\u003c/sup\u003e (R-squared/coefficient of determination); (2) Q\u003csup\u003e2\u003c/sup\u003e (The predictive relevance); (3)path coefficient effect size; (4) SRMR (Standardized Mean Root Square Residual); (5) VIF(Variance Inflation Factor); (6) BIC (Bayesian Information Criterion). First, potential multicollinearity in the structural model was examined by calculating the Variance inflation factor (VIF) values, and it was found that all values were below 3.30 (Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e), well below the recommended threshold of \u0026lt; 5 (Hair et al., \u003cspan class=\"CitationRef\"\u003e2017\u003c/span\u003e). These results confirm that multicollinearity has not reached a critical level among the predictor constructs in the model. Furthermore, the cross-validated redundancy (Q\u003csup\u003e2\u003c/sup\u003e) is greater than zero, indicating that the model has predictive relevance.\u003c/p\u003e \u003cp\u003eThe direct path coefficients of the structural model were analyzed to examine the hypothesized relationships between variables. Based on a sample of 142 respondents, the following findings were observed. The analysis of the coefficients in the TAM constructed in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e revealed that the path from CA to PEU was small but statistically significant (\u003cem\u003eβ\u003c/em\u003e = 0.321, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01), supporting H3. The path from SFS to PEU was also significant (\u003cem\u003eβ\u003c/em\u003e = 0.374, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01), supporting H5. In addition, the path from TS to PEU was significant as well (\u003cem\u003eβ\u003c/em\u003e = 0.176, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05), providing support for H7. Regarding PU, the path from CA to PU was significant (\u003cem\u003eβ\u003c/em\u003e = 0.291, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01), supporting H4, and the path from SFS to PU was also significant (\u003cem\u003eβ\u003c/em\u003e = 0.406, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01), supporting H6. However, the path from TS to PU was not statistically significant (\u003cem\u003eβ\u003c/em\u003e = 0.133, \u003cem\u003ep\u003c/em\u003e \u0026gt; 0.05), leading to the rejection of H8. As for ITU, the path from PEU to ITU showed a small positive effect but was not significant (\u003cem\u003eβ\u003c/em\u003e = 0.191, \u003cem\u003ep\u003c/em\u003e \u0026gt; 0.05), so H1 is not supported. In contrast, the path from PU to ITU showed a strong and statistically significant positive effect (\u003cem\u003eβ\u003c/em\u003e = 0.550, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001), thus H2 is supported. For a complete representation of the loadings, see Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e and Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003ctable id=\"Tab3\" border=\"1\"\u003e \u003ccaption\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSaturated model results\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003c/colgroup\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003e\u003cem\u003eR\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eAdj.\u003cem\u003eR\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eQ\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eRMSE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eMAE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eSRMR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eVIF\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eBIC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eChi-square\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePEU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.541\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.531\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.488\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.731\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"3\"\u003e \u003cp\u003e0.071\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"3\"\u003e \u003cp\u003e\u0026lt; 3.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e-91.845\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" rowspan=\"3\"\u003e \u003cp\u003e1066.211\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.506\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.495\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.439\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.772\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.536\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e-81.213\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eITU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.522\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.515\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.441\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.762\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.599\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e-90.826\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003ctable id=\"Tab4\" border=\"1\"\u003e \u003ccaption\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePath coefficients and hypothesis testing\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003c/colgroup\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\"\u003e \u003cp\u003eHypothesis\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eRelationships\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eStandard deviation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003et-statistics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eResults\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eH1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePEU -\u0026gt; ITU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.189\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.137\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e1.382\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eUnsupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eH2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePU -\u0026gt; ITU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.552\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.132\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e4.185\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eH3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA -\u0026gt;PEU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.318\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.114\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e2.977\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eH4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA -\u0026gt; PU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.296\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.117\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e2.722\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eH5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFS -\u0026gt; PEU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.382\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.132\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e2.772\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eH6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFS -\u0026gt; PU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.404\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.141\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e2.707\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eH7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTS -\u0026gt; PEU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.077\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e2.225\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.026\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eH8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTS -\u0026gt; PU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.133\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.072\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e1.815\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eUnsupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eQualitative Comparative Analysis of Fuzzy Sets (fsQCA)\u003c/h2\u003e \u003cdiv id=\"Sec17\" class=\"Section3\"\u003e \u003ch2\u003eSelection and calibration of variables\u003c/h2\u003e \u003cp\u003eBased on the literature and empirical research findings, CA, SFS, TS, PEU, and PU are selected as condition variables, while ITU serves as the outcome variable. Following the approach of Rihoux and Ragin (\u003cspan class=\"CitationRef\"\u003e2009\u003c/span\u003e), the 95th percentile, 50th percentile, and 5th percentile of the data are set as calibration anchors, representing full membership, the crossover point, and full non-membership, respectively. The calibration process is carried out using fsQCA 4.1 software.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eNecessity analysis of conditions\u003c/h2\u003e \u003cp\u003eBefore conducting the configurational analysis, the necessity of each condition variable (including its negation) is tested to determine whether it constitutes a necessary condition for ITU. This is done through an analysis of the necessity of each individual condition variable for the outcome variable (as shown in the Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e). The results indicate that the consistency of all condition variables is less than 0.9, meaning that none of them are necessary conditions for a high/low ITU. Therefore, further configurational analysis is required.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" class=\"colspec\"\u003e\u003c/div\u003e\u003ctable id=\"Tab5\" border=\"1\"\u003e \u003ccaption\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDetails of the necessity analyses\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003c/colgroup\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" rowspan=\"2\"\u003e \u003cp\u003eCondition Variables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\"\u003e \u003cp\u003eITU\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\"\u003e \u003cp\u003e~ITU\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\"\u003e \u003cp\u003eConsistency\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eCoverage\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eConsistency\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003eCoverage\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.826\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.831\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.649\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.587\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e~CA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.590\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.590\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.814\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.808\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.878\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.799\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.694\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.568\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e~SFS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.525\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.656\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.754\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.847\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.773\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.763\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.653\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.580\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e~TS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.574\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.648\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.733\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.744\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePEU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.809\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.844\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.621\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.583\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e~PEU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.600\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.638\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.834\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.797\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.858\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.675\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.589\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e~PU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.576\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.663\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.808\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\"\u003e \u003cp\u003e0.837\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003csup\u003eNote: “~”denotes the set relationship of “negation” or “complement”\u003c/sup\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eSufficiency analysis of condition configurations\u003c/h2\u003e \u003cp\u003eFollowing the approach of Pappas \u0026amp; Woodside (\u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e) and Leppänen (2023), for medium-sized samples (50 \u0026lt; \u003cem\u003eN\u003c/em\u003e \u0026lt; 150), the frequency threshold is set at 2, and the raw consistency threshold is set at 0.8. After constructing the truth table, Standard Analyses are performed. Given the numerous factors influencing teachers’ ITU in the research context, and the difficulty in measuring the necessity of each factor, the condition variables are uniformly classified as “Present or Absent”. In QCA studies, intermediate solutions, which are reasonable, well-founded, and of moderate complexity, are typically the preferred choice for reporting and interpretation. Therefore, the analysis primarily focuses on the intermediate solution, with simple solutions used as supplementary, ultimately identifying the core and peripheral conditions influencing the outcome variable. Under the Intermediate Solution type, eight configurational paths are identified, but the results are not concise. After adjusting the frequency threshold to 3 and performing Standard Analyses again, five configurational paths are obtained. The reduction in the number of paths decreases redundancy between them, making the analysis clearer. While the Solution Coverage of these five paths slightly decreases (from 0.878 to 0.836), it remains at a high level; however, Solution Consistency improves (from 0.788 to 0.850), surpassing the acceptable minimum threshold of 0.8, indicating that the logic of the paths is more robust. The adjustment of the frequency threshold is thus deemed appropriate. Therefore, the five configurational paths presented in Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e can be regarded as the sufficient conditions for influencing teachers' intention to continue using the system in the given research context.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\"\u003e\u003c/div\u003e\u003ctable id=\"Tab6\" border=\"1\"\u003e \u003ccaption\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSufficiency Analysis of Condition Configurations\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003c/colgroup\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" rowspan=\"2\"\u003e \u003cp\u003eCondition Variables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"5\"\u003e \u003cp\u003eITU\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eCA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e•\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSFS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e⊙\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e⊙\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e⊙\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eTS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e⊙\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e⊙\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePEU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e⊙\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e⊙\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003ePU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e•\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e×\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e•\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eConsistency\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.918\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.907\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.860\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.915\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.937\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eRaw coverage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.664\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.744\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.398\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.294\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.325\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eUnique coverage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.026\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.084\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003eSolution consistency\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"5\"\u003e \u003cp\u003e0.850\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\"\u003e \u003cp\u003esolution coverage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"5\"\u003e \u003cp\u003e0.836\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003csup\u003eNote: “•” denotes peripheral conditions, “×” denotes non−existent conditions (negated conditions), “⊙” denotes core conditions, and a blank space indicates conditions that are optional\u003c/sup\u003e.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003eConfiguration 1 (CA*SFS*TS) indicates that, in the process of teachers using the artificial intelligence ethics learning chatbot, CA serves as a peripheral condition providing support, while SFS and TS act as core conditions playing a critical role. This suggests that when the system functions are stable and technical support is adequate, the accuracy of the content enhances teachers’ trust in the chatbot and their intention to use it. Therefore, Configuration 1 can be labeled as “Technological quality satisfaction type”. The consistency of this configuration is 0.918, with a raw coverage of 0.664 and unique coverage of 0.026. This means that this combinational path can explain 66.4% of the cases related to teachers’ intention to accept the chatbot, and 2.6% of the cases can be exclusively explained by this path.\u003c/p\u003e \u003cp\u003eConfiguration 2 (SFS*PEU*PU) suggests that, during the use of the artificial intelligence ethics learning chatbot, SFS and PEU play core roles, while PU serves as a peripheral condition offering supplementary support. This implies that when the chatbot’s system functionality is stable and teachers find the chatbot easy to use, although perceived usefulness is not a decisive factor, it still helps to some extent in enhancing teachers’ intention to use the chatbot. Thus, Configuration 2 can be named “System experience satisfaction type”. The consistency of this configuration is 0.906, with a raw coverage of 0.744 and unique coverage of 0.084. This indicates that the combinational path can explain 74.4% of the cases regarding teachers’ intention to accept the chatbot, and 8.4% of the cases can be independently explained by this path.\u003c/p\u003e \u003cp\u003eConfiguration 3 (~ CA*SFS*~PEU*~PU) shows that, in the process of using the artificial intelligence ethics learning chatbot, SFS serves as the sole core condition playing a crucial role. This suggests that when the chatbot’s system functionality is stable, the absence of other conditions does not significantly affect teachers’ intention to use the chatbot. The stability of the system itself provides teachers with confidence and assurance to continue using the chatbot. Therefore, Configuration 3 can be labeled as “Functionality stability satisfaction type”. The consistency of this configuration is 0.860, with a raw coverage of 0.397 and unique coverage of 0.023. This means that this combinational path can explain 39.7% of the cases related to teachers’ intention to accept the chatbot, and 2.3% of the cases can be independently explained by this path.\u003c/p\u003e \u003cp\u003eConfiguration 4 (~ CA*~SFS*~TS*PEU*~PU) indicates that, in the process of using the chatbot, PEU serves as the sole core condition playing a key role. This suggests that when teachers perceive the chatbot as easy to operate, the absence of other conditions does not significantly influence their intention to use it. The chatbot’s ease of use alone is sufficient to motivate teachers to continue using it. Thus, Configuration 4 can be labeled as “Operational convenience satisfaction type”. The consistency of this configuration is 0.914, with a raw coverage of 0.293 and unique coverage of 0.008. This implies that this combinational path can explain 29.3% of the cases related to teachers’ intention to accept the chatbot, and 0.8% of the cases can be exclusively explained by this path.\u003c/p\u003e \u003cp\u003eConfiguration 5 (~ CA*~SFS*TS*~PEU*PU) suggests that, during the use of the chatbot, SFS acts as a core condition, while PU serves as a peripheral condition providing auxiliary support. This implies that when the system functionality is stable, teachers’ perceived usefulness of the chatbot will further enhance their intention to use it. Therefore, Configuration 5 can be labeled as “System value satisfaction type”. The consistency of this configuration is 0.936, with a raw coverage of 0.325 and unique coverage of 0.002. This indicates that the combinational path can explain 32.5% of the cases related to teachers’ intention to accept the chatbot, and 0.2% of the cases can be explained by this path.\u003c/p\u003e \u003c/div\u003e "},{"header":"Conclusion, implications and limitations","content":"\u003cp\u003eThis study integrates the TAM with a mixed-methods approach, combining PLS-SEM and fsQCA, to uncover the diverse factors influencing teachers’ intention to continue using the chatbot. Based on data from 142 participants, PLS-SEM results indicate that perceived usefulness significantly and positively impacts behavioral intention, while perceived ease of use does not show a significant direct effect. Further fsQCA analysis identifies five sufficient configurations leading to high intention to continue use: Technological quality satisfaction, system experience satisfaction, functionality stability satisfaction, operational convenience satisfaction, and system value satisfaction. These findings highlight the multidimensional nature of technology acceptance in AI ethics education and offer empirical evidence and practical implications for the design of effective and AI-supported learning tools.\u003c/p\u003e\u003cp\u003eThe results of the PLS-SEM analysis showed that perceived usefulness had a significant positive impact on teachers’ behavioral intention to use the chatbot. This finding is consistent with previous research on continued use intention of information systems (Daneji et al., \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e; Huang, \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e). Foroughi et al. (2023) further pointed out that in educational contexts, perceived usefulness is an important driving force for teachers to adopt AI tools, especially when the chatbot can effectively support ethical reasoning and teaching practices. The results indicate that when teachers perceive the chatbot’s value in improving their understanding and application of AI ethics, they are more likely to continue using it. In contrast, perceived ease of use had a nonsignificant direct effect on intention to use, which may be attributed to teachers’ high overall digital literacy and therefore their lower sensitivity to usability barriers (Amhag et al., \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e; Saxena \u0026amp; Doleck, 2023). Ngo et al. (\u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e) further point out that in scenarios with high cognitive load, functional value is often more influential than ease of use in technology adoption. In terms of external factors, content accuracy and system function significantly improve perceived usefulness and perceived ease of use, indicating that information reliability and technical performance are the basis for shaping user value perception. For example, in AI-driven assistance systems, teachers are often more willing to tolerate a certain level of operational complexity as long as the system has substantial benefits in teaching (Gârdan et al., 2025). This result aligns with the research of Fan and Jiang (2024), who found similar patterns in their study of AI-assisted learning tools. Conversely, while technical support has a positive effect on perceived ease of use, it has no significant effect on perceived usefulness. This may be because teachers pay more attention to the chatbot’s professional capabilities and autonomous problem-solving characteristics rather than relying on external support services.\u003c/p\u003e\u003cp\u003eThis study has several limitations that should be acknowledged. First, the relatively small sample size may limit the representativeness and diversity of the findings, making it difficult to capture the full range of teachers’ perceptions and acceptance across different educational levels, subject backgrounds, or institutional settings. Second, the LLM-based chatbot used in this study is still in its early developmental stage, with system instability and incomplete functionality—such as interface issues and delayed responses—which may have interfered with participants’ experiences and affected their actual acceptance and usage behaviors. Third, as most participants were from economically developed regions, the sample may carry regional and cultural biases, reducing the extent to which the findings reflect the perspectives of educators in more diverse socioeconomic and cultural contexts. Fourth, although the study employed a mixed-methods approach incorporating both PLS-SEM and fsQCA, the qualitative interpretation derived from fsQCA configurations may lack depth due to the absence of complementary narrative data (e.g., interviews or open-ended responses). Without richer contextual insights from participants, the configurational findings—though informative—may not fully capture the underlying rationale or motivational reasoning behind teachers’ behavioral intentions.\u003c/p\u003e\u003cp\u003eBuilding on the insights of this study, future research should consider several avenues for further exploration. First, expanding the sample to include a larger and more demographically diverse population, particularly teachers from underrepresented or rural regions, could provide a more nuanced and generalizable understanding of the factors influencing acceptance of AI ethics tools. Second, further technical development of the chatbot is essential. Future studies could assess the impact of specific chatbot improvements, such as adaptive learning features, multimodal interaction, or real-time feedback, on user satisfaction and learning outcomes. Third, cross-cultural or cross-regional comparative studies would be valuable in uncovering how contextual variables, such as cultural attitudes toward AI, policy environments, or institutional readiness, influence ethical AI adoption in education. Longitudinal designs could also offer insights into the sustained impact and evolving perceptions of AI-integrated teaching tools over time. Fourth, to enhance the explanatory power of fsQCA findings, future studies could incorporate qualitative interviews or case studies to triangulate results and provide deeper, context-rich interpretations of teachers’ acceptance patterns and ethical reasoning.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthical Approval Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research protocol was approved by the Research Ethics Committee of the Graduate Institute, Wenzhou University (approval number: WZU-2025-0930C, approval date: September 30, 2025), and all procedures comply with the requirements of the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed consent statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll subjects signed a written informed consent form between October 15 and October 22, 2025, and voluntarily participated after fully understanding the purpose, procedures, risks and rights of the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated during this study are fully available within the article and supplementary materials, which include the post-test questionnaire data on teachers\u0026apos; usage of the LLM-based chatbot. The supplementary files provide the raw dataset with basic coding applied.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eChenchen Liu contributed to the Resources, Supervision, Funding acquisition, and Writing - Review \u0026amp; Editing. Yangying Yi contributed to the Conceptualization, Methodology, Formal analysis, Writing - Original Draft, and Writing - Review \u0026amp; Editing. Xinying Wang contributed to the Investigation, Data Curation, Writing - Original Draft, and Writing - Review \u0026amp; Editing. Haijie Wang contributed to the Conceptualization, Writing - Original Draft, and Writing - Review \u0026amp; Editing. Youmei Wang contributed to the Resources, Supervision, Project administration, and Writing - Review \u0026amp; Editing.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAl-Adwan, A. S., Li, N., Al-Adwan, A., Abbasi, G. A., Albelbisi, N. A., \u0026amp; Habibi, A. (2023). Extending the technology acceptance model (TAM) to Predict University Students\u0026rsquo; intentions to use metaverse-based learning platforms.\u003cem\u003e Education and Information Technologies, 28\u003c/em\u003e(11), 15381-15413. https://doi.org/ 10.1007/s10639-023-11816-3\u003c/li\u003e\n\u003cli\u003eAl-Adwan, A. S., Nofal, M., Akram, H., Albelbisi, N. A., \u0026amp; Al-Okaily, M. (2022). Towards a Sustainable Adoption of E-Learning Systems: The Role of Self-Directed Learning. \u003cem\u003eJournal of Information Technology Education: Research, 21\u003c/em\u003e. https://doi.org/10.28945/4980\u003c/li\u003e\n\u003cli\u003eAl-Emran, M., Mezhuyev, V., \u0026amp; Kamaludin, A. (2018). Technology Acceptance Model in M-learning context: A systematic review. \u003cem\u003eComputers \u0026amp; Education, 125\u003c/em\u003e, 389-412. https:// doi.org/10.1016/j.compedu.2018.06.008\u003c/li\u003e\n\u003cli\u003eAljuaid, H. (2024). The impact of artificial intelligence tools on academic writing instruction in higher education: A systematic review. \u003cem\u003eArab World English Journal (AWEJ) Special Issue on ChatGPT\u003c/em\u003e, 26-55. https://dx.doi.org/10.24093/ awej/ChatGPT.2\u003c/li\u003e\n\u003cli\u003eAlmaiah, M. A., \u0026amp; Al Mulhem, A. (2019). Analysis of the essential factors affecting of intention to use of mobile learning applications: A comparison between universities adopters and non-adopters. \u003cem\u003eEducation and Information\u003c/em\u003e\u003cem\u003e \u003c/em\u003e\u003cem\u003eTechnologies,24\u003c/em\u003e(2),1433-1468.https://doi.org/10.1007/ s10639-018-9840-1\u003c/li\u003e\n\u003cli\u003eAlmaiah, M. A., \u0026amp; Alamri, M. M. (2018). Proposing a new technical quality requirements for mobile learning applications. \u003cem\u003eJournal of Theoretical and Applied Information Technology, 96\u003c/em\u003e(19), 6955-6968. \u003c/li\u003e\n\u003cli\u003eAlmaiah, M. A., Jalil, M. A., \u0026amp; Man, M. (2016). Extending the TAM to examine the effects of quality features on mobile learning acceptance. \u003cem\u003eJournal of Computers in Education, 3\u003c/em\u003e, 453-485.https://doi.org/10.1007/s40692-016-0074-1\u003c/li\u003e\n\u003cli\u003eAlmuhanna, M. A. (2024). Teachers\u0026rsquo; perspectives of integrating AI-powered technologies in K-12 education for creating customized learning materials and resources. \u003cem\u003eEducation and Information Technologies\u003c/em\u003e, 1-29.https://doi.org/ 10.1007/s10639-024-13257-y\u003c/li\u003e\n\u003cli\u003eAmhag, L., Hellstr\u0026ouml;m, L., \u0026amp; Stigmar, M. (2019). Teacher Educators\u0026rsquo; Use of Digital Tools and Needs for Digital Competence in Higher Education. \u003cem\u003eJournal of Digital Learning in Teacher Education, 35\u003c/em\u003e(4), 203\u0026ndash;220. https://doi.org/10.1080/21532974.2019.1646169\u003c/li\u003e\n\u003cli\u003eArteaga S\u0026aacute;nchez, R., Duarte Hueros, A., \u0026amp; Garc\u0026iacute;a Ordaz, M. (2013). E‐learning and the University of Huelva: a study of WebCT and the technological acceptance model. \u003cem\u003eCampus-Wide Information Systems, 30\u003c/em\u003e(2), 135-160. https://doi.org/10.1108/ 10650741311306318\u003c/li\u003e\n\u003cli\u003eAyanwale, M. A., Adelana, O. P., Molefi, R. R., Adeeko, O., \u0026amp; Ishola, A. M. (2024). Examining artificial intelligence literacy among pre-service teachers for future classrooms. \u003cem\u003eComputers and education open, 6\u003c/em\u003e(1)..https://doi.org/10.1016/j.caeo. 2024.100179\u003c/li\u003e\n\u003cli\u003eBaek, T. H., \u0026amp; Kim, M. (2023). Is chatgpt scary good? How user motivations affect creepiness and trust in generative artificial intelligence. \u003cem\u003eTelematics and Informatics, 83\u003c/em\u003e(3). https://doi.org/ 10.1016/j.tele.2023.102030\u003c/li\u003e\n\u003cli\u003eBai, B., Wang, J., \u0026amp; Chai, C. S. (2021). Understanding Hong Kong primary school English teachers\u0026rsquo; continuance intention to teach with ICT.\u003cem\u003e Computer Assisted Language Learning, 34\u003c/em\u003e(4), 528-551. https://doi.org/10.1080/09588221.2019. 1627459\u003c/li\u003e\n\u003cli\u003eBakar-Corez, A., \u0026amp; Kocaman-Karoglu, A. (2024). E-dishonesty among postgraduate students and its relation to self-esteem. \u003cem\u003eEducation and Information Technologies, 29\u003c/em\u003e(7), 8275-8300. https:// doi.org/10.1007/s10639-023-12105-9\u003c/li\u003e\n\u003cli\u003eBarrett, A. J., Pack, A., \u0026amp; Quaid, E. D. (2021). Understanding learners\u0026rsquo; acceptance of high-immersion virtual reality systems: Insights from confirmatory and exploratory PLS-SEM analyses.\u003cem\u003e Computers \u0026amp; Education, 169\u003c/em\u003e(81). https:// doi.org/10.1016/j.compedu.2021.104214\u003c/li\u003e\n\u003cli\u003eBelda-Medina, J., \u0026amp; Calvo-Ferrer, J. R. (2022). Using Chatbots as AI Conversational Partners in Language Learning. \u003cem\u003eApplied Sciences, 12\u003c/em\u003e(17), 8427. https://doi.org/ 10.3390/app12178427\u003c/li\u003e\n\u003cli\u003eBozkurt, A. (2023). Generative artificial intelligence (AI) powered conversational educational agents: The inevitable paradigm shift.\u003cem\u003e Asian Journal of Distance Education, 18\u003c/em\u003e(1), 198-204.https:// doi.org/10.5281/zenodo.7716416\u003c/li\u003e\n\u003cli\u003eCheng, Y. M. (2012). Effects of quality antecedents on e‐learning acceptance. \u003cem\u003eInternet Research, 22\u003c/em\u003e(3), 361-390.https://doi.org/10.1108/10662241211235699\u003c/li\u003e\n\u003cli\u003eChin, W. W. (1998). The partial least squares approach to structural equation modeling. \u003cem\u003eModern methods for business research, 295\u003c/em\u003e(2), 295-336.\u003c/li\u003e\n\u003cli\u003eDaneji, A. A., Ayub, A. F. M., \u0026amp; Khambari, M. N. M. (2019). The effects of perceived usefulness, confirmation and satisfaction on continuance intention in using massive open online course (MOOC). \u003cem\u003eKnowledge Management \u0026amp; E-Learning, 11\u003c/em\u003e(2), 201-214.https://doi.org/10.34105/ j.kmel.2019.11.010\u003c/li\u003e\n\u003cli\u003eDavis, F. D., Bagozzi, R. P., \u0026amp; Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. \u003cem\u003eManagement science, 35\u003c/em\u003e(8), 982-1003.https://doi.org/ 10.1287/mnsc.35.8.982\u003c/li\u003e\n\u003cli\u003eDeLone, W. H., \u0026amp; McLean, E. R. (1992). Information systems success: The quest for the dependent variable.\u003cem\u003e Information Systems Research, 38\u003c/em\u003e(6), 373\u0026ndash;384.https:// doi.org/10.1287/isre.3.1.60\u003c/li\u003e\n\u003cli\u003eDeLone, W. H., \u0026amp; McLean, E. R. (2003). The DeLone and McLean model of information systems success: A 10-year update.\u003cem\u003e Journal of Management Information Systems, 19\u003c/em\u003e(4),9\u0026ndash;30.https:// doi.org/10.1080/07421222.2003.11045748\u003c/li\u003e\n\u003cli\u003eDu, W., Liang, R. Y., \u0026amp; Liu, D. (2022). Factors influencing school teachers\u0026rsquo; continuous usage intention of using VR technology for classroom teaching. \u003cem\u003eSage Open, 12\u003c/em\u003e(3). https://doi.org/ 10.1177/21582440221114325\u003c/li\u003e\n\u003cli\u003eDwivedi, Y. K., Wade, M. R., \u0026amp; Schneberger, S. L. (Eds.). (2011). Information systems theory: Explaining and predicting our digital society, vol. 1. https://doi.org/ 10.1007/978-1-4419-6108-2\u003c/li\u003e\n\u003cli\u003eEke, D. O. (2023). Chatgpt and the rise of generative AI: Threat to academic integrity? \u003cem\u003eJournal of Responsible Technology, 13\u003c/em\u003e.https://doi.org/10.1016/j.jrt. 2023.100060\u003c/li\u003e\n\u003cli\u003eFarhan, W., Razmak, J., Demers, S., \u0026amp; Laflamme, S. (2019). E-learning systems versus instructional communication tools: Developing and testing a new e-learning user interface from the perspectives of teachers and students. \u003cem\u003eTechnology in Society, 59\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e https://doi.org/10.1016/ j.techsoc.2019.101192. \u003c/li\u003e\n\u003cli\u003eFornell, C. \u0026amp; Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. \u003cem\u003eJournal of Marketing Research 18\u003c/em\u003e, 39-50.https://doi.org/ 10.1177/002224378101800104\u003c/li\u003e\n\u003cli\u003eGajic, N., \u0026amp; Boolaky, M. (2015). Impact of technical support on customer satisfaction: Case of automotive paints. \u003cem\u003eSAGE Open, 5\u003c/em\u003e(4).https://doi.org/10.1177/2158244 015611457\u003c/li\u003e\n\u003cli\u003eG\u0026acirc;rdan, I. P., Manu, M. B., G\u0026acirc;rdan, D. A., Negoiță, L. D. L., Paștiu, C. A., Ghiță, E., \u0026amp; Zaharia, A. (2025, February). Adopting AI in education: optimizing human resource management considering teacher perceptions. \u003cem\u003eIn Frontiers in Education, 10\u003c/em\u003e. https://doi.org/10.3389/feduc. 2025.1488147\u003c/li\u003e\n\u003cli\u003eGhavifekr, S., \u0026amp; Rosdy, W. A. W. (2015). Teaching and learning with technology: Effectiveness of ICT integration in schools.\u003cem\u003e International journal of research in education and science, 1\u003c/em\u003e(2), 175-191.https://doi.org/10.21890/ijres.23596\u003c/li\u003e\n\u003cli\u003eGiang, N. T. T., \u0026amp; Nga, P. T. H. (2024). The influence of information systems on student satisfaction: A study of perceived ease of use and perceived usefulness. \u003cem\u003eJournal of Financial-Marketing Research, 15\u003c/em\u003e(2), 62-72. https://doi.org/10.52932/jfm.vi2. 479\u003c/li\u003e\n\u003cli\u003eG\u0026oacute;mez-Ramirez, I., Valencia-Arias, A., \u0026amp; Duque, L. (2019). Approach to M-learning acceptance among university students: An integrated model of TPB and TAM. \u003cem\u003eInternational Review of research in open and distributed learning, 20\u003c/em\u003e(3).https://doi.org/10.19173/irrodl.v20i4.4061 CopiedAn error has occu\u003c/li\u003e\n\u003cli\u003eGr\u0026ouml;nroos, C. (1984). A service quality model and its marketing implications. \u003cem\u003eEuropean Journal of marketing, 18\u003c/em\u003e(4), 36-44.https://doi.org/10.1108/EUM0000000004784\u003c/li\u003e\n\u003cli\u003eHagendorff, T. (2020). The ethics of AI ethics: An evaluation of guidelines. \u003cem\u003eMinds and machines, 30\u003c/em\u003e(1), 99-120.https://doi.org/10.1007/s11023-020-09517-8\u003c/li\u003e\n\u003cli\u003eHair, J. F., Hult, G. T. M., Ringle, C. M., \u0026amp; Sarstedt, M. (2017). A primer on partial least squaresstructural equation modeling (PLS-SEM).\u003cem\u003e International Journal of Research \u0026amp; Method in Education, 38\u003c/em\u003e(2), 220\u0026ndash;221.https://doi.org/10.1080/ 1743727X.2015.1005806\u003c/li\u003e\n\u003cli\u003eHan, J. H., \u0026amp; Sa, H. J. (2022). Acceptance of and satisfaction with online educational classes through the technology acceptance model (TAM): The COVID-19 situation in Korea. \u003cem\u003eAsia Pacific Education Review, 23\u003c/em\u003e(3), 403-415.https:// doi.org/10.1007/s12564-021-09716-7\u003c/li\u003e\n\u003cli\u003eHsu, L. (2016). Examining EFL teachers\u0026rsquo; technological pedagogical content knowledge and the adoption of mobile-assisted language learning: a partial least square approach.\u003cem\u003e Computer Assisted Language Learning, 29\u003c/em\u003e(8), 1287\u0026ndash;1297. https://doi.org/10.1080/09588221.2016.1278024\u003c/li\u003e\n\u003cli\u003eHu, H. H., Kandampully, J., \u0026amp; Juwaheer, T. D. (2009). Relationships and impacts of service quality, perceived value, customer satisfaction, and image: an empirical study. \u003cem\u003eThe service industries journal, 29\u003c/em\u003e(2), 111-125.https://doi.org/10.1080/ 02642060802292932\u003c/li\u003e\n\u003cli\u003eHu, Y. H. (2022). Effects and acceptance of precision education in an AI-supported smart learning environment. \u003cem\u003eEducation and Information\u003c/em\u003e\u003cem\u003e \u003c/em\u003e\u003cem\u003eTechnologies,27\u003c/em\u003e(2),2013-2037.https://doi.org/10.1007/ s10639-021-10664-3\u003c/li\u003e\n\u003cli\u003eHuang, Y. M. (2019). Examining students\u0026apos; continued use of desktop services: Perspectives from expectation-confirmation and social influence.\u003cem\u003e Computers in Human Behavior, 96\u003c/em\u003e, 23-31.https:// doi.org/10.1016/j.chb.2019.02.010\u003c/li\u003e\n\u003cli\u003eIbrahim, A., \u0026amp; Shiring, E. (2022). The Relationship between Educators\u0026apos; Attitudes, Perceived Usefulness, and Perceived Ease of Use of Instructional and Web-Based Technologies: Implications from Technology Acceptance Model (TAM). \u003cem\u003eInternational Journal of Technology in Education, 5\u003c/em\u003e(4), 535-551.https:// doi.org/10.46328/ijte.285\u003c/li\u003e\n\u003cli\u003eIpek, Z. H., Gozum, A. C., Papadakis, S., \u0026amp; Kallogianakis, M. (2023). Educational applications of the ChatGPT AI system: A systematic review research. Educational Process.\u003cem\u003e International Journal, 12\u003c/em\u003e(3), 26\u0026ndash;55.https://doi.org/ 10.22521/edupij.2023.123.2\u003c/li\u003e\n\u003cli\u003eJobin, A., Ienca, M., \u0026amp; Vayena, E. (2019). The global landscape of AI ethics guidelines. \u003cem\u003eNature machine intelligence, 1\u003c/em\u003e(9), 389-399. https://doi.org/10.1038/s42256-019-0088-2\u003c/li\u003e\n\u003cli\u003eKalankesh, L. R., Nasiry, Z., Fein, R. A., \u0026amp; Damanabi, S. (2020). Factors influencing user satisfaction with information systems: a systematic review. \u003cem\u003eGalen Medical Journal, 9\u003c/em\u003e. https:// doi.org/10.31661/gmj.v9i0.1686.\u003c/li\u003e\n\u003cli\u003eKasneci, E., Se\u0026szlig;ler, K., K\u0026uuml;chemann, S., Bannert, M., Dementieva, D., Fischer, F., ... \u0026amp; Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for education.\u003cem\u003e Learning and individual differences, 103\u003c/em\u003e.https://doi.org/10.1016/j.lindif.2023.102274\u003c/li\u003e\n\u003cli\u003eKazim, E., \u0026amp; Koshiyama, A. S. (2021). A high-level overview of AI ethics. \u003cem\u003ePatterns, 2\u003c/em\u003e(9). https://doi.org/10.1016/j.patter.2021.100314\u003c/li\u003e\n\u003cli\u003eKe, C. H., Sun, H. M., \u0026amp; Yang, Y. C. (2012). Effects of User and System Characteristics on Perceived Usefulness and Perceived Ease of Use of the Web-Based Classroom Response System. \u003cem\u003eTurkish Online Journal of Educational Technology, 11\u003c/em\u003e(3), 128-143.https://files.eric.ed.gov/ fulltext/EJ989206.pdf\u003c/li\u003e\n\u003cli\u003eKim, N. J., \u0026amp; Kim, M. K. (2022, March). Teacher\u0026rsquo;s perceptions of using an artificial intelligence-based educational tool for scientific writing. \u003cem\u003eIn Frontiers in education, 7\u003c/em\u003e. https://doi.org/10.3389/feduc.2022.755914\u003c/li\u003e\n\u003cli\u003eKim, Y., \u0026amp; Lee, H. S. (2014). Quality, perceived usefulness, user satisfaction, and intention to use: An empirical study of ubiquitous personal robot service. \u003cem\u003eAsian Social Science, 10\u003c/em\u003e(11), 1.https://doi.org/10.5539/ass.v10n11p1 \u003c/li\u003e\n\u003cli\u003eKing, W., \u0026amp; He, J. (2006). A meta-analysis of the technology acceptance model. \u003cem\u003eInformation \u0026amp; Management, 43\u003c/em\u003e, 740-755.https://doi.org/10.1016/j.im.2006. 05.003 \u003c/li\u003e\n\u003cli\u003eLepp\u0026auml;nen, P., George, G., \u0026amp; Alexy, O. (2023). When do novel business models lead to high performance? A configurational approach to value drivers, competitive strategy, and firm environment. \u003cem\u003eAcademy of management journal, 66\u003c/em\u003e(1), 164-194.https://doi.org/10.5465/ amj.2020.0969\u003c/li\u003e\n\u003cli\u003eLim, W. M., Gunasekara, A., Pallant, J. L., Pallant, J. I., \u0026amp; Pechenkina, E. (2023). Generative AI and the future of education: Ragnar\u0026ouml;k or reformation? A paradoxical perspective from management educators. \u003cem\u003eThe international journal of management education, 21\u003c/em\u003e(2),.https://doi.org/10.1016/ j.ijme.2023.100790\u003c/li\u003e\n\u003cli\u003eLiu, I. F., Chen, M. C., Sun, Y. S., Wible, D., \u0026amp; Kuo, C. H. (2010). Extending the TAM model to explore the factors that affect intention to use an online learning community. \u003cem\u003eComputers \u0026amp; education, 54\u003c/em\u003e(2), 600-610. https://doi.org/10.1016/ j.compedu.2009.09.009\u003c/li\u003e\n\u003cli\u003eMa, Q., \u0026amp; Liu, L. (2004). The technology acceptance model: A meta-analysis of empirical findings. \u003cem\u003eJournal of Organizational and End User Computing, 16\u003c/em\u003e(1), 59-72.https://doi.org/10.4018/ 978-1-59140-474-3.ch006\u003c/li\u003e\n\u003cli\u003eMachdar, N. M. (2016). The effect of information quality on perceived usefulness and perceived ease of use. \u003cem\u003eBusiness and entrepreneurial review, 15\u003c/em\u003e(2), 131-146. https://doi.org/10.25105/ ber.v15i2. 4630\u003c/li\u003e\n\u003cli\u003eMahande, R. D., Jasruddin, J., \u0026amp; Nasir, N. (2019). Is success model for EDMODO e-learning user satisfaction through TAM on students. \u003cem\u003eJournal of Educational Science and Technology, 5\u003c/em\u003e(2), 140-152.https://doi.org/10.26858/est.v5i2.9575\u003c/li\u003e\n\u003cli\u003eMouta, A., Torrecilla-S\u0026aacute;nchez, E. M., \u0026amp; Pinto-Llorente, A. M. (2025). Comprehensive professional learning for teacher agency in addressing ethical challenges of AIED: Insights from educational design research. \u003cem\u003eEducation and Information Technologies, 30\u003c/em\u003e(3), 3343-3387.https://doi.org/ 10.1007/s10639-024-12946-y\u003c/li\u003e\n\u003cli\u003eNelson, R. R., Todd, P. A., \u0026amp; Wixom, B. H. (2005). Antecedents of information and system quality: an empirical examination within the context of data warehousing.\u003cem\u003e Journal of management information systems, 21\u003c/em\u003e(4), 199-235. https://doi.org/ 10.1080/07421222.2005.11045823\u003c/li\u003e\n\u003cli\u003eNgo, T. T. A., Tran, T. T., An, G. K., \u0026amp; Nguyen, P. T. (2024). ChatGPT for educational purposes: Investigating the impact of knowledge management factors on student satisfaction and continuous usage. \u003cem\u003eIEEE Transactions on Learning Technologies, 17\u003c/em\u003e, 1367\u0026ndash;1378.https://doi.org/ 10.1109/TLT.2024.3383773\u003c/li\u003e\n\u003cli\u003eNguyen, A., Ngo, H. N., Hong, Y., Dang, B., \u0026amp; Nguyen, B. P. T. (2023). Ethical principles for artificial intelligence in education. \u003cem\u003eEducation and information technologies, 28\u003c/em\u003e(4), 4221-4241.\u003c/li\u003e\n\u003cli\u003eNikolopoulou, K. (2021). Mobile Devices and Mobile Learning in Greek Secondary Education: Policy, Empirical Findings and Implications. \u003cem\u003eIn Handbook for online learning contexts: Digital, mobile and open: Policy and practice \u003c/em\u003e(pp. 67-80). https://doi.org/10.1007/978-3-030-67349-9_6\u003c/li\u003e\n\u003cli\u003eOhliati, J., \u0026amp; Abbas, B. S. (2019). Measuring students satisfaction in using learning management system. \u003cem\u003eInternational Journal of Emerging Technologies in Learning (Online), 14\u003c/em\u003e(4), 180-189. https://doi.org/10.3991/ijet.v14i04.9427\u003c/li\u003e\n\u003cli\u003eOnwuegbuzie, A. J., \u0026amp; Daniel, L. G. (2002). A framework for reporting and interpreting internal consistency reliability estimates. \u003cem\u003eMeasurement and evaluation in counseling and development, 35\u003c/em\u003e(2), 89-103. https://doi.org/10.1080/07481756. 2002.12069052\u003c/li\u003e\n\u003cli\u003ePadayachee, I. (2017). Educator perceptions of virtual learning system quality characteristics. \u003cem\u003eSouth African Computer Journal, 29\u003c/em\u003e(3), 95-126. https://hdl. handle.net/10520/EJC-c5c8388c1\u003c/li\u003e\n\u003cli\u003ePappas, I. O., \u0026amp; Woodside, A. G. (2021). Fuzzy-set Qualitative Comparative Analysis (fsQCA): Guidelines for research practice in Information Systems and marketing. \u003cem\u003eInternational Journal of Information Management, 58\u003c/em\u003e. https://doi.org/10.1016/ j.ijinfomgt.2021.102310\u003c/li\u003e\n\u003cli\u003ePatra, I., Alazemi, A., Al-Jamal, D., \u0026amp; Gheisari, A. (2022). The effectiveness of teachers\u0026rsquo; written and verbal corrective feedback (CF) during formative assessment (FA) on male language learners\u0026rsquo; academic anxiety (AA), academic performance (AP), and attitude toward learning (ATL). \u003cem\u003eLanguage Testing in Asia, 12\u003c/em\u003e(1), 19. https://doi.org/10.1186/s40468-022-00169-2\u003c/li\u003e\n\u003cli\u003eRai, A., Lang, S. S., \u0026amp; Welker, R. B. (2002). Assessing the validity of IS success models: An empirical test and theoretical analysis. \u003cem\u003eInformation systems research, 13\u003c/em\u003e(1), 50-69.https://doi.org/ 10.1287/isre.13.1.50.96\u003c/li\u003e\n\u003cli\u003eRihoux, B., \u0026amp; Ragin, C. C. (2009). Configurational comparative methods: Qualitative comparative analysis (QCA) and related techniques. \u003cem\u003eSAGE Publications, Inc\u003c/em\u003e.https://doi.org/10.4135/ 9781452226569\u003c/li\u003e\n\u003cli\u003eScherer, R., Siddiq, F., \u0026amp; Teo, T. (2015). Becoming more specific: Measuring and modeling teachers\u0026apos; perceived usefulness of ICT in the context of teaching and learning. \u003cem\u003eComputers \u0026amp; Education, 88\u003c/em\u003e, 202-214. https://doi.org/10.1016/j. compedu.2015.05.005\u003c/li\u003e\n\u003cli\u003eScherer, R., Siddiq, F., \u0026amp; Tondeur, J. (2019). The technology acceptance model (TAM): A meta-analytic structural equation modeling approach to explaining teachers\u0026rsquo; adoption of digital technology in education. \u003cem\u003eComputers \u0026amp; education, 128\u003c/em\u003e, 13-35.https://doi.org/10.1016/ j.compedu.2018.09.009\u003c/li\u003e\n\u003cli\u003eSeddon, P. B. (1997). A respecification and extension of the DeLone and McLean model of IS success.\u003cem\u003e Information systems research, 8\u003c/em\u003e(3), 240-253. https://doi.org/ 10.1287/isre.8.3.240\u003c/li\u003e\n\u003cli\u003eShen, X., \u0026amp; Liu, J. (2022). Analysis of factors affecting user willingness to use virtual online education platforms.\u003cem\u003e International Journal of Emerging Technologies in Learning, 17\u003c/em\u003e(1), 74-89. https://doi.org/10.3991/ijet.v17i01.28713\u003c/li\u003e\n\u003cli\u003eSiafis, V., Rangoussi, M., \u0026amp; Psaromiligkos, Y. (2024). Recommender Systems for Teachers: A Systematic Literature Review of Recent (2011\u0026ndash;2023) Research. \u003cem\u003eEducation Sciences, 14\u003c/em\u003e(7), 723.https://doi.org/10.3390/educsci14070723\u003c/li\u003e\n\u003cli\u003eTeo, T. (2010). A path analysis of pre-service teachers\u0026apos; attitudes to computer use: applying and extending the technology acceptance model in an educational context. \u003cem\u003eInteractive Learning Environments, 18\u003c/em\u003e(1), 65-79. https://doi.org/10. 1080/10494820802231327\u003c/li\u003e\n\u003cli\u003eUddin, M.M. (2024). Rejection or integration of AI in academia: determining the best choice through the Opportunity Cost theoretical formula. \u003cem\u003eDiscover Education\u003c/em\u003e\u003cem\u003e,\u003c/em\u003e\u003cem\u003e 3\u003c/em\u003e(249).https://doi.org/10.1007/ s44217-024-00349-7\u003c/li\u003e\n\u003cli\u003eWaheed, M., \u0026amp; Jam, F. A. (2010). Teacher\u0026rsquo;s intention to accept online education: Extended TAM model. \u003cem\u003eInterdisciplinary Journal of Contemporary Research in Business, 2\u003c/em\u003e(5), 330-344.\u003c/li\u003e\n\u003cli\u003eWang, S., Sun, Z., \u0026amp; Chen, Y. (2023). Effects of higher education institutes\u0026rsquo; artificial intelligence capability on students\u0026rsquo; self-efficacy, creativity and learning performance. \u003cem\u003eEducation and Information Technologies, 28\u003c/em\u003e(5), 4919\u0026ndash;4939. https://doi.org/10.1007/s10639-022-11338-4.\u003c/li\u003e\n\u003cli\u003eWang, Y., Liu, C., \u0026amp; Tu, Y. F. (2021). Factors affecting the adoption of AI-based applications in higher education. \u003cem\u003eEducational Technology \u0026amp; Society, 24\u003c/em\u003e(3), 116-129.https://www.jstor.org/ stable/27032860\u003c/li\u003e\n\u003cli\u003eWiese, L. J., Patil, I., Schiff, D. S., \u0026amp; Magana, A. J. (2025). AI Ethics Education: A Systematic Literature Review. \u003cem\u003eComputers and Education: Artificial Intelligence, 8\u003c/em\u003e. https://doi.org/10.1016/ j.caeai.2025.100405\u003c/li\u003e\n\u003cli\u003eXu, J., Benbasat, I., \u0026amp; Cenfetelli, R. T. (2013). Integrating service quality with system and information quality: An empirical test in the e-service context. \u003cem\u003eMIS quarterly, 37\u003c/em\u003e(3), 777-794. https://www.jstor.org/stable/43825999\u003c/li\u003e\n\u003cli\u003eZhang, C., Schie\u0026szlig;l, J., Pl\u0026ouml;\u0026szlig;l, L., Hofmann, F., \u0026amp; Gl\u0026auml;ser-Zikuda, M. (2023). Acceptance of artificial intelligence among pre-service teachers: a multigroup analysis. \u003cem\u003eInternational Journal of Educational Technology in Higher Education, 20\u003c/em\u003e(1), 49.https://doi.org/10.1186/s41239- 023-00420-7\u003c/li\u003e\n\u003cli\u003eZheng, Y., Zhao, K., \u0026amp; Stylianou, A. (2013). The impacts of information quality and system quality on users\u0026apos; continuance intention in information-exchange virtual communities: An empirical investigation. \u003cem\u003eDecision support systems, 56\u003c/em\u003e, 513-524.https://doi.org/10.1016/j.dss.2012.11.008\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"AI ethics, PLS-SEM, fsQCA, Technology acceptance model","lastPublishedDoi":"10.21203/rs.3.rs-8618094/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8618094/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAs Artificial Intelligence (AI) becomes increasingly and widely integrated in education, fostering teachers\u0026rsquo; understanding of AI ethics has become critical. This study investigates the acceptance of a chatbot based on a Large Language Model (LLM), designed to enhance teachers\u0026rsquo; understanding of AI ethics in education. A mixed-methods approach combining Partial Least Squares Structural Equation Modeling (PLS-SEM) and fuzzy-set Qualitative Comparative Analysis (fsQCA) was applied in the current study with 142 participants. Findings show that the content accuracy, system functionality and stability significantly influence perceived ease of use and perceived usefulness. FsQCA further revealed five sufficient configurations associated with high continuance intention. These findings provide practical guidance for designing AI-supported learning environments and offer directions for future research.\u003c/p\u003e","manuscriptTitle":"Understanding AI ethics in education with a LLM-based chatbot: Evidence from PLS-SEM and fsQCA analyses","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-30 19:18:27","doi":"10.21203/rs.3.rs-8618094/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-07T09:11:43+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-02T17:25:47+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"124311870625038980995030629100001910848","date":"2026-04-25T11:23:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"289751673686085786549638984042024326765","date":"2026-04-09T07:48:54+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-05T04:47:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"233175751576015087738001601350878651930","date":"2026-03-27T04:27:16+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-16T12:49:18+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-16T12:31:20+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-02-18T09:00:24+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-15T12:07:18+00:00","index":"","fulltext":""},{"type":"submitted","content":"Humanities and Social Sciences Communications","date":"2026-02-15T12:02:40+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"0ce051a9-b8ce-4c15-9982-b219f0e74911","owner":[],"postedDate":"March 30th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-07T09:11:43+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-02T17:25:47+00:00","index":160,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[{"id":64582329,"name":"Social science/Education"},{"id":64582330,"name":"Business and commerce/Information systems and information technology"},{"id":64582331,"name":"Physical sciences/Mathematics and computing"},{"id":64582332,"name":"Social science/Science technology and society"}],"tags":[],"updatedAt":"2026-05-07T09:24:32+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-30 19:18:27","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8618094","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8618094","identity":"rs-8618094","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}