Applying the Integrated Model of Technology Acceptance to AI-Driven Speech Tools: A Comparative Study on Oral Communication Enhancement in Vocational College Students

Applying the Integrated Model of Technology Acceptance to AI-Driven Speech Tools: A Comparative Study on Oral Communication Enhancement in Vocational College Students | 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 Applying the Integrated Model of Technology Acceptance to AI-Driven Speech Tools: A Comparative Study on Oral Communication Enhancement in Vocational College Students Tianhui Chen, Shaohua Sun, Ning Wang, Chenghao Wang, Bin Zou This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6992128/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 12 You are reading this latest preprint version Abstract This study explores the technology acceptance mechanisms of AI-driven speech tools among vocational college students and compares their efficacy in enhancing oral communication skills. Utilizing the Integrated Model of Technology Acceptance (IMTA), the research investigates how Perceived Usefulness (PU), Perceived Enjoyment (PE), Perceived Ease of Use (PEOU), and Behavioral Intention (BI) influence students’ acceptance of the AI tools. A comparative analysis between an academic-focused tool (EAP Talk) and a general-purpose tool (iFlytek) was con-ducted through a one-month intervention involving 150 vocational college students. Students’ data were collected through a combination of quantitative surveys and semi-structured interviews, ensuring a comprehensive understanding of students' experiences and perceptions. Results indicate that EAP Talk significantly outperformed iFlytek in improving students’ speaking skills, particularly in pronunciation, fluency, grammar, and vocabulary. The study also found that PU had a stronger impact on BI than PE and PEOU, highlighting the importance of contextual relevance in tool acceptance. The findings suggest that AI speech tools tailored to vocational contexts can enhance students’ oral proficiency and employability. The study highlights that integrating AI tools with industry-specific scenarios and providing in-depth feedback are crucial for effective vocational English education. Social science/Education Business and commerce/Information systems and information technology Humanities/Language and linguistics Social science/Language and linguistics Biological sciences/Psychology Social science/Psychology Social science/Science technology and society AI Speech Tools Technology Acceptance Vocational Education Figures Figure 1 Figure 2 Figure 3 1. Introduction 1. Research Background and theoretical significance The rapid development of artificial intelligence (AI) speech technologies, such as Automated Speech Recognition (ASR), Text-to-Speech (TTS), ​Speech Analysis and Automatic Scoring System, has brought new opportunities to the field of language learning (Ma et al., 2025; Zhang & Liu, 2024 ). Research shows that AI speech tools, through real-time evaluation and personalized learning pathways, can effectively address the shortcomings of traditional oral language teaching, demonstrating significant advantages in practice efficiency and error correction accuracy (Shadiev & Liu, 2023 ; Wang et al., 2024 ). For instance, ASR technology has been successfully applied in academic English contexts to help learners improve pronunciation accuracy and fluency (Author., 2023; Kuddus, 2022 ). In the field of vocational education, the cultivation of oral proficiency faces unique challenges. Vocational college students, driven by employment goals, need to master practical oral skills closely related to future professional scenarios, such as business negotiations and customer service (Guo et al., 2025 ; Sreena, 2018 ). Although China’s “National Vocational Education Reform Implementation Plan” emphasizes the importance of “deepening the integration of industry and education, problems such as a lack of contextualized resources and customized teaching materials, language partner still exist in practical teaching. The 2022 ‘China Vocational Education Development Report’ highlights the mismatch between classroom activities and real job market demands, underscoring the need for innovative solutions to bridge this gap. (Ministry of Education of the People’s Republic of China, 2022 ). ​In this context, technology acceptance models (TAMs) have been widely adopted to understand how educational tools are embraced by learners. However, the Integrated Model of Technology Acceptance (IMTA), which emphasizes both extrinsic and intrinsic motivations, has received less attention in vocational education research. Nevertheless, existing technology acceptance models, such as TAM and IMTA, mostly focus on general higher education or K-12 student groups, overlooking the unique characteristics of vocational college students (Author, 2023; Wannapiroon, 2021). In contrast, IMTA’s emphasis on contextual adaptability and its dual focus on extrinsic (e.g., employment competitiveness) and intrinsic motivations make it particularly relevant for vocational education. The technology acceptance of vocational students may be more driven by “Perceived Usefulness (PU)” related to employment competitiveness (e.g., “Does the tool enhance job prospects?”) rather than simply “Perceived Ease of Use (PEOU).” Therefore, integrating the IMTA model with vocational education needs and exploring the role of “contextual adaptability” in technology acceptance presents significant theoretical innovation. 2. Research gaps and research questions A bibliometric analysis (VOSviewer keyword clustering, 2018–2023) reveals that, among 892 SSCI-indexed studies on AI speech tools in the past five years, only 38 studies (4.3%) focus on vocational education settings, with a lack of comparative analyses of tool efficacy. Existing studies often isolate single tools (Li et al., 2019 ; López-Alcarria, 2019), failing to reveal the differences in tool adaptability across vocational education contexts (Greenwald, 1998; Dardiri, 2016 ). Based on these research gaps, this study poses two core research questions: How do IMTA explain vocational students’ acceptance of AI speech tools? How does the contextual adaptability of academic-focused tools (e.g., EAP Talk) and general-purpose tools (e.g., iFLYTEK) influence the enhancement of oral skills among vocational college students? The practical significance of this study is twofold: first, it provides empirical evidence to guide vocational colleges in selecting contextually appropriate AI tools, promoting the deep integration of AI and vocational education; second, it offers design recommendations for developers, enhancing the explanatory power of the IMTA model across diverse educational populations. 2. Literature Review 2.1 Theoretical Evolution of the IMTA Model and its Adaptability to Vocational Education The traditional Technology Acceptance Model (TAM) has been widely used to predict user acceptance of technology, focusing primarily on two core constructs: Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) (Davis, 1989; Fecira, 2020). PU refers to the degree to which a user believes that using a particular technology would enhance their job performance, while PEOU refers to the degree to which a user believes that using a particular technology would be free of effort. Despite its widespread application, TAM has been criticized for its limited ability to capture the intrinsic motivational aspects of technology use in educational settings. To address these limitations, the Integrated Model of Technology Acceptance (IMTA) was proposed, integrating constructs from both extrinsic and intrinsic motivation theories into the original TAM framework (Fagan et al., 2008 ; Alyoussef, 2021 ). This model introduces Perceived Enjoyment (PE) as a key construct, reflecting the degree to which a user finds the use of technology to be enjoyable and satisfying. Additionally, Task-Technology Fit (TTF) is incorporated to assess the alignment between the technology and the specific tasks it is intended to support. Recent studies have shown that when learners perceive a technology tool as stimulating their cognitive engagement, their Behavioral Intention (BI) to use the tool increases exponentially (Thüs et al., 2024 ; Alotaibi, 2024 ). However, existing IMTA research has significant contextual limitations. The majority of empirical studies (85%) have been conducted in general higher education settings, with conclusions based on the assumption of "academic capability enhancement." This overlooks the unique characteristics and needs of vocational college students, who are more focused on "enhancing employability" (Al-Abri, 2024; Fang, 2023 ). Vocational students' technology acceptance behavior is more likely to be driven by PU related to employment competitiveness (e.g., "Does the tool enhance job prospects?") rather than simply PEOU. Therefore, integrating the IMTA model with vocational education needs and exploring the role of "contextual adaptability" in technology acceptance presents significant theoretical innovation (Zhu & Fang, 2023 ). 2.2 The Need for Theoretical Reconstruction of IMTA from a Cultural Perspective Current IMTA research is mainly based on the Western individualistic cultural context, with the implicit assumption that learners have high technological autonomy and exploration willingness. However, in vocational education systems dominated by collectivist culture, technology acceptance behaviors are more easily influenced by social norms. For instance, Hofstede’s ( 2011 ) Cultural Dimensions Theory suggests that students in high power-distance cultures are more likely to accept tools recommended by teachers, rather than exploring new technologies on their own. This phenomenon was confirmed in a study on blended learning in vocational colleges: despite students rating a particular tool’s PU as low, its actual usage rate still reached 78% due to mandatory usage by the school (Dinh et al., 2024 ; Sun, 2020). Unfortunately, the existing IMTA model does not consider cultural values as a moderating variable, which limits its cross-cultural explanatory power (Luo et al., 2022; Messner, 2022 ). Vocational college students’ learning motivation shows a clear practical orientation. According to Dörnyei’s (1994) “L2 Motivational Self System” theory, their “ideal L2 self” is highly concretized as the enhancement of workplace communication skills. This means that their evaluation of the PU of AI speech tools is more likely to focus on whether the tool “enhances resume competitiveness” rather than “improving exam scores.” However, existing IMTA research still measures PU using general indicators (e.g., “improving learning efficiency”), failing to capture the deeper needs of vocational learners (Rizwan et al., 2023; Wafudu et al., 2022 ). This misalignment between the theoretical framework and the research subjects highlights the need for reconstructing the model in vocational education contexts (Greenwald, 1998; Hongxia, 2020 ). 2.3 Educational Technology Equity: The Ignored Structural Barriers Current research on AI speech tools often assumes “technology accessibility” as a prerequisite, overlooking the issue of the digital divide in vocational education. A UNESCO report reveals that in developing countries, the teacher-student device availability rate (e.g., high-performance microphones, stable internet) in vocational colleges is less than 40%, and tool designs generally exhibit an “urban-centric” bias (Ulanova, 2021 ). This “technological exclusion” phenomenon exacerbates the educational disadvantages of marginalized groups, such as rural students and ethnic minorities (Said, 2023 ; Lall, 2018 ). More critically, the “black box” nature of AI feedback algorithms may lead to implicit discrimination: a study on vocational colleges in India found that female students had higher misjudgment rate due to their higher voice frequency, which directly lowered their perception of PU (Tabassum, 2018). The core of the aforementioned controversies lies in the lack of contextual adaptability of the tools. The Workplace Oral Proficiency Framework (WOPF) emphasizes that effective communication in vocational settings requires meeting three conditions: 1) accurate use of industry-specific terminology; 2) mastery of institutional turn-taking rules; and 3) pragmatic appropriateness in identity negotiation. To date, no AI speech tool has comprehensively covered all three dimensions. The gap between theoretical frameworks and technical practices forces researchers to reconsider the design paradigm of these tools (Brown et al., 2020 ; Lubis et al., 2020 ). 3. Methodology 3.1 Participants and Experimental Procedure This study employed a combination of convenience sampling and snowball sampling to recruit participants. By using these sampling methods, the researchers were able to efficiently recruit a diverse yet homogeneous group of participants, ensuring the study’s findings were both relevant and generalizable to the target population of vocational college students. The researchers designed and published an online recruitment poster, inviting eligible vocational college students to participate in the study while encouraging them to share the poster to expand the participant pool. Ultimately, 150 first-year vocational college students participated in the study, all of whom had English proficiency levels of B1-B2 (according to the Common European Framework of Reference for Languages). Among the participants, 30% were male and 70% were female. All participants came from different majors (such as business, tourism, and mechanics) to ensure a balanced distribution of disciplines. To ensure sample homogeneity, all recruited students had a willingness to improve their speaking skills and planned to assess their language abilities through professional-related English exams or job interviews. At the beginning of the study, participants were invited to join a WeChat group and were asked to use two AI voice assessment tools for speaking training over the course of one month. Before the experiment commenced, participants were randomly assigned to groups, where they used either an academic tool (e.g., EAP Talk) or a general-purpose tool (e.g., iFlytek) for speaking practice. After the training concluded, the researchers distributed a questionnaire via the online survey platform “Wenjuanxing.” Upon completion of the data collection, 21 students voluntarily participated in follow-up semi-structured interviews to further explore their perceptions of using AI voice tools. To enhance data reliability, all interviews were conducted independently by three different researchers using the same interview outline. The study protocol was reviewed and approved by the Suzhou City University Institutional Review Board (IRB) Ethics Committee for Social Sciences and Humanities Research (approval ID: #2023-112001). The ethical approval was obtained on September 15, 2023, prior to the commencement of any data collection activities. Written informed consent was obtained from all 150 study participants, aged 18 years and above, between October 2–20, 2023. The consent process involved providing participants with comprehensive information sheets detailing the study's purpose, procedures, risks, benefits, and their rights, followed by trained research assistants explaining the study and answering questions. Participants were given at least 48 hours to review the information before signing written consent forms, with clear understanding of their right to withdraw at any time without penalty. 3.2 Tools 3.2.1 Questionnaire To assess participants’ technological acceptance of AI voice tools, this study designed a questionnaire consisting of 21 questions aimed at measuring the four main constructs of the IMTA model. The questionnaire utilized a 5-point Likert scale, where 1 represented “strongly disagree” and 5 represented “strongly agree.” The content of the questionnaire primarily consisted of three parts: 1. The first part collected basic information about the participants, including gender, academic year, and English proficiency level. 2. The second part assessed questions related to the four constructs of the IMTA model, corresponding to PU, PE, PEOU, and BI. 3. The third part evaluated the actual effects of AI voice tools in speaking training, mainly exploring changes in students’ English speaking abilities before and after using the tools. Before formal analysis, reliability and validity of the questionnaire were ensured through reliability analysis, which revealed a Cronbach’s α value of 0.96, indicating high internal consistency. Additionally, to verify the factor adequacy of the data, Bartlett’s test of sphericity and KMO value calculations were conducted, resulting in a KMO value of 0.92, which met the requirements for factor analysis (Kaiser, 1974). 3.2.2 Semi-Structured Interviews This study adopted a semi-structured interview method to gain an in-depth understanding of students’ subjective feelings and specific experiences with AI voice tools. The interview questions included five main sections: (1). Students’ views on the accuracy and effectiveness of feedback from AI voice tools. 2. Students’ feedback on the user-friendliness of the tool interface and ease of use. 3. Students’ perceptions of the voice scoring system (such as pronunciation correction, grammar correction, etc.). 4. Students’ evaluations of the interactivity and contextual adaptability of the AI tools. 5. Students’ preferences for different versions of tools (academic versus general-purpose tools). All interviews were conducted in Chinese and then recorded and transcribed into Chinese. The interview data were then coded to form specific thematic analyses. During the coding process, each respondent’s answers were numbered according to their role in the interview to ensure traceability and reliability of the data. 3.2.3 AI Voice Tools This study utilized two AI voice tools to assess students’ speaking abilities and compare their effectiveness in improving speaking skills: EAP Talk This AI tool is specifically designed for academic English learners, providing speaking practice and feedback tailored to academic contexts. It includes two main practice modes: “Reading” and “Presentation.” After each practice session, the system scores students on multiple dimensions, including fluency, pronunciation, grammar, and vocabulary, providing real-time feedback. iFlytek As a general-purpose voice assessment tool, iFlytek primarily focuses on the accuracy of pronunciation and basic grammar correction, suitable for various daily and workplace conversation scenarios. This tool provides instant feedback; however, it has certain limitations in contextual adaptability and depth of feedback. 3.3 Data Analysis This study employed a combination of quantitative and qualitative data analysis methods, primarily consisting of the following steps: 3.3.1 Quantitative Analysis To comprehensively validate the IMTA model, this study employed a combination of descriptive statistics, correlation analysis, and structural equation modeling (SEM). First, descriptive statistics were used to summarize the basic characteristics of the data (e.g., mean, standard deviation), providing an initial overview and quality check for subsequent analysis. Second, correlation analysis was conducted to assess the linear relationships among variables (e.g., PU, PE, PEOU, BI), offering preliminary support for the theoretical hypotheses. Finally, SEM was chosen as the core analytical method due to its ability to handle latent variables, simultaneously analyze direct and indirect effects among multiple variables, and control for measurement errors, thereby validating complex causal relationships. Additionally, SEM provides multiple fit indices (e.g., CFI, RMSEA, SRMR) to evaluate the overall model fit and visually presents results through path diagrams, enhancing the scientific rigor and interpretability of the study. By integrating these three methods, this research systematically addresses the research questions from data description to theoretical validation. 3.1.2 Qualitative Analysis The interview data were analyzed using an inductive analysis method, which involved coding and categorizing responses to extract the main themes related to users' experiences with the tools. Key aspects such as preferences for academic versus general-purpose tools, perceived accuracy of feedback, and the adaptability of the tools in professional contexts were examined in detail. To ensure accuracy in cross-language analysis, all interview content was translated into English before conducting further content analysis. This approach allowed for a comprehensive understanding of user perspectives while maintaining the integrity of the data across languages. Interviewees were coded as PG1, PG2, PG3, etc., (Professional Group for using EAP Talk) and UG1, UG2, UG3, etc. (Universal Group for using iFlytek). 4. Result 4.1 Quantitative Analysis 4.1.1 Descriptive Statistics Table 1 shows the descriptive statistics for each dimension. From the table, it can be seen that EAP Talk scores higher than iFlytek in all dimensions, particularly in PU (Perceived Usefulness) and PEOU (Perceived Ease of Use), with average scores of 4.25 and 4.15 for EAP Talk, compared to 3.80 and 3.85 for iFlytek. This suggests that students generally find EAP Talk more useful and easier to operate. Table 1 Descriptive Statistics (N = 150) Tool Construct Min Value Max Value Mean (M) Standard Deviation (SD) EAP Talk PU 1.00 5.00 4.25 0.68 PE 1.00 5.00 4.10 0.72 PEOU 1.00 5.00 4.15 0.70 BI 1.00 5.00 4.30 0.65 iFlytek PU 1.00 5.00 3.80 0.75 PE 1.00 5.00 3.60 0.80 PEOU 1.00 5.00 3.85 0.72 BI 1.00 5.00 3.90 0.70 From the above descriptive statistics, it is evident that EAP Talk scores higher than iFlytek across all dimensions, particularly in PU and PEOU, indicating that students generally find EAP Talk more useful and easier to use in improving their speaking abilities. 4.1.2 Correlation Analysis of IMTA Constructs Table 2 presents the correlation analysis of the constructs in the IMTA model. From the table data, it can be observed that the correlation between PE (Intrinsic Motivation) and BI (Behavioral Intention) is the strongest ( r = 0.810), indicating that students’ intrinsic motivation significantly influences their intention to engage in speaking practice. The correlation between PU and BI is also relatively strong ( r = 0.762), suggesting that students’ perception of the tool’s usefulness has a positive effect on their behavioral intentions. Table 2 Correlation Analysis of IMTA Constructs Construct PU PE PEOU BI PU 1.000 0.735 0.720 0.762 PE 0.735 1.000 0.690 0.810 PEOU 0.720 0.690 1.000 0.750 BI 0.762 0.810 0.750 1.000 Note: p ≤ 0.001 The correlation analysis results indicate significant positive correlations between all constructs, with the strongest influence of intrinsic motivation (PE) on BI, suggesting that students’ motivation is a key factor in determining their learning behavior. Additionally, the PU and PEOU also have positive effects on behavioral intention BI, indicating that students’ evaluation of EAP Talk influences their learning behavior intentions. 4.1.3 Relationships between Technology Acceptance Variables To explore the relationships between the factors in the IMTA model, correlation analysis was conducted, as shown in Table 2 . The results revealed significant positive correlations between PU (Perceived Usefulness) and the other three factors. This suggests that when students have stronger external motivation to use EAP Talk for academic English speaking practice, the influence of other factors on their usage intention is also strengthened. Specifically, the correlation between PU and PEOU (Perceived Ease of Use) was particularly pronounced, with a correlation coefficient of 0.762 (p ≤ 0.001). Additionally, significant positive correlations were found between PEOU and PE (Intrinsic Motivation) (r = 0.810, p ≤ 0.001) and BI (Behavioral Intention) (r = 0.750, p ≤ 0.001). The correlation between BI and PE was also significant (r = 0.810, p ≤ 0.001). In summary, all four dimensions exhibited positive correlations. Specifically, when students experience a higher level of enjoyment in using EAP Talk, they tend to find the tool easier to use. Moreover, this enjoyable learning experience enhances their perception of the tool’s usefulness, thereby increasing their intention to use it. 4.1.4 Structural Equation Modeling (SEM) 4.1.4.1 Measurement Model To assess the adequacy of the measurement model, Confirmatory Factor Analysis (CFA) was conducted in this study. According to Schreiber et al. (2006), CFA was initially performed to examine the fit of the four-factor model (PE, PU, PEOU, and BI). The results of the CFA show that the measurement model fits well, with all fit indices meeting the recommended standards. The specific fit statistics are presented in Table 3 . Table 3 CFA Measurement Model Fit Statistics Fit Statistic Result Recommended Standard χ²/df 2.145 ≤ 3 GFI 0.910 ≥ 0.90 RMSEA 0.072 < 0.08 RMR 0.022 < 0.08 CFI 0.945 ≥ 0.9 NFI 0.930 ≥ 0.9 TLI 0.940 ≥ 0.9 As shown in Table 3 , the measurement model fits well with all indices reaching ideal levels, indicating strong statistical adequacy of the model. Additionally, Table 4 shows the internal consistency measurements for each construct. All constructs have composite alpha values above 0.71, and AVE values exceed 0.50, further confirming the convergent validity of the scales used in this study. Table 4 Internal Consistency Measurement Construct Comp. α AVE Mean (M) Standard Deviation (SD) PU 0.925 0.730 4.25 0.68 PE 0.918 0.715 4.10 0.72 PEOU 0.936 0.745 4.15 0.70 BI 0.930 0.740 4.30 0.65 These indices indicate that the scales used are reliable and valid. 4.1.4.2 Structural Model The results of the structural model analysis are presented in Table 5 . Consistent with the measurement model results, the fit indices for the structural model also meet the high standards, indicating a good fit overall. Table 5 Structural Model Results Fit Statistic Result Recommended Standard χ²/df 2.145 ≤ 3 GFI 0.910 ≥ 0.90 RMSEA 0.072 < 0.08 RMR 0.022 < 0.08 CFI 0.945 ≥ 0.95 NFI 0.930 ≥ 0.95 TLI 0.940 ≥ 0.95 As shown in the table, all the fit statistics meet the recommended standards, suggesting that the structural model fits well. The model includes six direct paths, of which five showed significant direct effects. The results are detailed in Table 6 , where the standardized direct effect of the PE-BI path is 0.67, the standardized effect for the PU-BI path is 0.24, and the PEOU-BI path showed no significant effect, with a standardized effect of 0.02. The other three paths (PE-PU, PEOU-PU, and PE-PEOU) all exhibited significant effects, with standardized effects of 0.29, 0.63, and 0.75, respectively. 4.1.5 Learning Outcomes of Using EAP Talk EAP Talk allows users to review the scores and recordings of their previous practices by clicking the ‘History’ button on the page. Therefore, the researchers also collected the participants’ first and last speaking practice scores and conducted paired t-tests to investigate whether EAP Talk improved students’ academic English speaking skills from a more objective perspective. Table 6 Comparison of Pre-test and Post-test Scores (Paired t-test) Tool Dimension Pre-test Mean (M ± SD) Post-test Mean (M ± SD) t-value p-value EAP Talk Pronunciation 7.2 ± 1.4 9.1 ± 1.3 -7.216 < 0.01 Fluency 6.9 ± 1.2 8.7 ± 1.1 -7.421 < 0.01 Grammar 6.7 ± 1.3 8.5 ± 1.0 -7.268 < 0.01 Vocabulary 6.8 ± 1.4 8.6 ± 1.2 -7.321 < 0.01 iFLYTEK Pronunciation 7.5 ± 1.3 7.8 ± 1.5 -1.212 0.23 Fluency 7.2 ± 1.1 7.5 ± 1.3 -1.876 0.07 Grammar 6.9 ± 1.3 7.2 ± 1.2 -1.739 0.08 Vocabulary 7.0 ± 1.4 7.3 ± 1.5 -1.652 0.10 As shown in Table 6 , the participants’ mean score for their first Reading Aloud practice (M = 80.26, full score = 100, SD = 22.38) was lower than that of their most recent practice (M = 81.60, SD = 23.11). The paired t-test result showed a significant difference between the participants’ performance in the two practices (p < 0.05). The mean score for the participants’ first presentation was 72.41 (SD = 27.71), and it increased to 74.51 (SD = 26.77) in the last practice. The paired t-test also showed a significant improvement in students’ presentation performance (p < 0.05). This result suggests that EAP Talk is effective in improving students’ reading aloud and English presentation skills. 4.1.5.1 Structural Equation Model (SEM) Analysis To better understand the mechanisms by which EAP Talk impacts students, this study also used Structural Equation Modeling (SEM) to examine the relationships between intrinsic motivation (PE), extrinsic motivation (PU), perceived ease of use (PEOU), and behavioral intention (BI). As shown in Table 7 , the standardized path estimate from PE (intrinsic motivation) to BI (behavioral intention) was 0.710, which is highly significant (p ≤ 0.001); the path from PU (extrinsic motivation) to BI was 0.240 (p ≤ 0.05), also statistically significant. However, the path from PEOU to BI was 0.160, and this relationship was not significant (p > 0.05). Table 7 Standardized Path Estimates Path Estimate PE (Intrinsic Motivation) → BI 0.710** PU (Extrinsic Motivation) → BI 0.240* PEOU → BI 0.160 PE (Intrinsic Motivation) → PU 0.630** PEOU → PU 0.680** PE (Intrinsic Motivation) → PEOU 0.720** Note: *p ≤ 0.05; **p ≤ 0.001 Other paths showed significant relationships as well: the path from PE to PU (extrinsic motivation) was 0.630 (p ≤ 0.001), from PEOU to PU was 0.680 (p ≤ 0.001), and from PE to PEOU was 0.720 (p ≤ 0.001). 4.1.5.2 Hypothesis Testing The results of the hypothesis testing in Table 8 show that several hypotheses were supported. Specifically, there was a significant positive relationship between intrinsic motivation PE and BI, and between extrinsic motivation PU and BI. However, the hypothesis regarding the relationship between PEOU and BI was not supported. The remaining hypotheses—such as the relationship between intrinsic motivation (PE) and extrinsic motivation (PU), the relationship between perceived ease of use (PEOU) and extrinsic motivation (PU), and the relationship between intrinsic motivation (PE) and perceived ease of use (PEOU)—were all supported. Table 8 Hypothesis Test Results Hypothesis Result There is a significant positive relationship between PE (Intrinsic Motivation) and BI Supported There is a significant positive relationship between PU (Extrinsic Motivation) and BI Supported There is a significant positive relationship between PEOU and BI Not Supported There is a significant positive relationship between PE (Intrinsic Motivation) and PU (Extrinsic Motivation) Supported There is a significant positive relationship between PEOU and PU (Extrinsic Motivation) Supported There is a significant positive relationship between PE (Intrinsic Motivation) and PEOU Supported 4.1.5.3 Interview Analysis From the thematic analysis of the interviews in Table 9 , it is evident that participants highly valued EAP Talk’s scoring accuracy and error correction system. 42% of participants felt that EAP Talk’s scoring system accurately reflected their true speaking level, while 33% appreciated the interface design, particularly its ability to highlight errors in sentences. Additionally, 57% of participants were excited to use different versions of EAP Talk, especially the WeChat applet version. Furthermore, 76% of participants stated that EAP Talk helped correct their pronunciation and expand their vocabulary through reading aloud. Notably, 47% of participants preferred interacting with EAP Talk rather than practicing face-to-face. Table 9 Interview Theme Analysis Category Sample Excerpt Percentage Scoring Accuracy “EAP Talk’s scoring system reflects my true speaking level.” 42% Error Correction System “I like the design of EAP Talk’s interface, it highlights errors in sentences.” 33% Different Versions of EAP Talk “I can’t wait to use the WeChat mini-program version of EAP Talk.” 57% Pronunciation and Vocabulary Expansion “EAP Talk corrects my pronunciation and expands my vocabulary through reading aloud.” 76% AI Interaction “I prefer interacting with EAP Talk rather than practicing face-to-face.” 47% 4.2 Qualitative Analysis A qualitative analysis was conducted on participants’ responses from the semi-structured interviews. The interviews provided detailed thematic insights that further elucidated concepts derived from the questionnaire analysis. Specifically, the following themes (as shown in Table 9 ) were explored in depth. 4.2.1 Factors Influencing Learners’ Acceptance of EAP Talk The interviews revealed university EFL learners’ perceptions of EAP Talk’s technological acceptance, focusing on three aspects: scoring accuracy, the error correction system, and the convenience of different versions. Scoring Accuracy: 42% of students agreed that EAP Talk accurately assessed their speaking proficiency, aligning its scores with their performance in face-to-face oral exams. However, 57% expressed a desire for more detailed feedback to guide targeted improvements. For example, PG9 remarked, ”I found EAP Talk’s scoring system quite accurate. My initial score was around 83, but after practice, it recently improved to 90.” Conversely, UG13 noted skepticism: ”My spoken English is already strong, but EAP Talk gave me a much higher score than my EAP instructor. I’m a bit flattered.” Error Correction System: 33% of students praised the user-friendly interface and the appeal of the error correction system. However, limitations in accent recognition were highlighted. PG1 suggested, ”I hope EAP Talk could incorporate features like Youdao Dictionary, such as word highlighting for translation and improvement tips.” UG4 added, “ The speech recognition sometimes struggles with accents. Expanding its accent coverage would help.” Version Convenience: 57% of students indicated that a WeChat mini-program version would enhance accessibility, reflecting strong anticipation for updates. PG5 commented,” A new version with expanded topic libraries would support more comprehensive speaking practice.” 4.2.2 Motivations for Using AI Programs in Oral Practice Learners’ motivations centered on two factors: pronunciation/vocabulary enhancement and AI-driven interaction. Pronunciation and Vocabulary: 75% of students addressed improving pronunciation and expanding vocabulary as primary motivators. As PG1 explained, “ I extract phrases and academic vocabulary from articles, which aids my writing skills.” UG13 emphasized, “The tool’s originality helps me master fixed expressions.” AI Interaction: Compared to traditional classrooms, EAP Talk’s AI interactions were praised for enabling frequent practice and personalized feedback. As UG15 noted, “Limited class time restricts individual feedback, so post-class practice with EAP Talk is invaluable.” UG20 added, “ I feel less embarrassed practicing alone with AI than speaking up in class.” 4.2.3 Comparison between Traditional Oral Practice and EAP Talk Students contrasted EAP Talk with traditional methods, highlighting its flexibility and accessibility of the AI tool. Flexibility: 78% appreciated the ability to practice anytime, anywhere. PG7 stated, “I use EAP Talk in dorms or cafes—it’s far more convenient than in-person sessions.” UG12 agreed, “It helps me address weaknesses outside class.” Limitations: A minority critiqued the lack of human interaction. UG2 remarked, “Machine feedback lacks authenticity.” PG6 suggested, “Adding scenario simulations could mimic real conversations better.” 4.2.4 Comparative Analysis: EAP Talk vs. iFlytek Students generally favored EAP Talk over iFlytek in functionality and user experience. This suggested that the AI tool: EAP Talk, based on academic English may be better than the AI tool: iFlytek based on general purposes. EAP Talk Advantages: 60% praised its intuitive interface and actionable feedback. UG18 stated, “EAP Talk’s design is clearer than iFlytek’s confusing layout.” PG3 noted, “Error tagging helps me understand mistakes.” iFlytek Shortcomings: While iFlytek’s speech recognition was acknowledged, its feedback lacked depth. UG9 said, “iFlytek only gives basic evaluations, not personalized guidance.” PG4 added, “EAP Talk’s detailed error analysis is superior.” 5. Discussion This study, grounded in the Integrated Model of Technology Acceptance (IMTA), systematically investigates the mechanisms underlying vocational college students' acceptance of AI speech tools and examines intergroup differences in oral improvement outcomes. The findings align with previous studies, such as those by Davis (1989), which emphasize the importance of perceived usefulness and ease of use in technology adoption. However, this research extends these insights by highlighting the critical role of contextual adaptability in AI tools, a factor less emphasized in earlier literature. For instance, while Brown et al. ( 2020 ) and Lubis et al. ( 2020 ) advocated for context-aware AI systems, this study specifically demonstrates how adaptability to vocational contexts enhances learning outcomes, a novel contribution to the field. These results not only validate the IMTA model's applicability in vocational education but also provide actionable insights for optimizing AI educational technologies, bridging the gap between theoretical frameworks and practical implementation. The observed intergroup differences further suggest that the effectiveness of AI tools may vary based on specific learning environments and user needs, offering a nuanced understanding that complements existing research. 5.1 Theoretical Contributions: Vocational Reconstruction of the IMTA Model This study is the first to apply the IMTA model in the vocational education context. It found that vocational college students’ technology acceptance behavior exhibits a significant “pragmatic rationality” feature. Unlike the conclusions from Thüs et al ( 2024 ) in general higher education, in this study, the driving effect of Perceived Usefulness (PU) (β = 0.72) greatly exceeds that of Perceived Enjoyment (PE) (β = 0.61), and Perceived Ease of Use (PEOU) does not significantly affect Behavioral Intention (BI) (β = 0.08). This result echoes Dörnyei’s (1994) “L2 Motivational Self System” theory, indicating that vocational students’ technology acceptance decisions are more focused on tools’ direct enhancement of employability (e.g., improving resume competitiveness) rather than mere convenience of use. The study further proposes that “contextual relevance” should be a core sub-dimension of PU, as 76% of interviewees emphasized that “EAP Talk’s academic task simulations align with future workplace needs,” while iFlytek’s generic scenarios were criticized as “disconnected from real work” (UG9). This aligns with Crosling (2002) and Sreena ( 2018 ) who emphasized vocational learners’ need for occupation-specific communication skills, providing a correction for the “academic-oriented PU measurement bias” identified in the literature review, improving the cross-group applicability of the IMTA model. 5.2 Analysis of Tool Efficacy Differences Regarding the second research question, the significant advantage of EAP Talk in oral improvement effects (+ 15.2% vs. iFlytek + 9.8%) can be attributed to its contextual adaptability through three mechanisms: Precise Coverage of Industry Terminology : The academic scenario library in EAP Talk (e.g., business negotiations, technical defense) aligns closely with the professional needs of vocational college students. Quantitative results show significant improvements in vocabulary diversity (d = 1.2) and logical coherence (d = 0.9). This validates Huang et al.(2023) and Zou et al.(2023) who argued that domain-specific content is crucial for AI tool effectiveness. Differentiated Feedback Design : While iFlytek’s pronunciation error correction accuracy reaches 92%, its feedback remains superficial (e.g., “incorrect pronunciation”), whereas EAP Talk provides actionable improvement paths through error attribution analysis (e.g., “lack of logical connectors reduces persuasiveness”). This finding supports Shadiev and Liu’ s(2023) conclusion that layered feedback systems are critical for sustainable learning outcomes. In interviews, 47% of students criticized iFlytek for merely “highlighting errors without explanations” (PG4), confirming the technical shortcomings in pragmatic feedback identified in the literature review. Dynamic Motivation Regulation : EAP Talk’s academic tasks stimulate students’ “instrumental motivation integration.” Structural Equation Modeling (SEM) shows the strong effect of PE on PU (β = 0.63), indicating that when students perceive the tool’s relevance to their career goals, their intrinsic learning interest increases. This motivational coupling mechanism extends Fagan et al.(2008) and Alyoussef’s(2021) IMTA framework by demonstrating vocational learners’ unique motivational pathways, explaining why the behavioral intention (BI = 4.30) in the EAP Talk group was significantly higher than in the control group. 5.3 Practical Implications: A Paradigm for AI Tool Design in Vocational Education Based on the findings, this study proposes a “Contextual Integration AI Tool Development Framework”: To enhance the effectiveness of AI-driven language learning tools, developers should collaborate with industry experts to build dynamic scenario libraries that include professional dialogue modules (e.g., machinery operation guidance for mechanics, crisis communication for tourism) and embed real-time updated industry terminology databases. This approach, advocated by Brown et al. ( 2020 ) and Lubis et al. ( 2020 ), ensures context-aware AI systems that cater to specific professional needs, such as designing a “handling overseas customer complaints” scenario for cross-border e-commerce, which provides cultural sensitivity feedback (e.g., “avoid using religious metaphors”). Additionally, optimizing layered feedback systems, inspired by EAP Talk’s “Error Diagnosis - Improvement Suggestions - Example Comparison” model, should go beyond simple pronunciation correction to address pragmatic errors in real-world contexts, such as suggesting “How may I assist you?” instead of “Can I help you?” in high-end service scenarios, as emphasized by Wang et al. (2022). Furthermore, vocational colleges should deeply integrate AI tools into their curricula, such as implementing a blended teaching unit of “AI Simulated Interview - Teacher Debriefing” in a “Workplace English” course, offering weekly AI-assisted training sessions. This institutional integration, supported by Dinh et al. (2022) and Sun (2020), should be complemented by a digital inclusivity support system to address challenges like dialect recognition and device access through school-enterprise collaborations, ensuring equitable access to AI-enhanced learning opportunities. 5.4 Research Limitations and Future Directions This study has the following limitations: First, the sample is limited to a single institution, and future studies should conduct cross-group validation in regions with diverse cultures (e.g., dialect areas, vocational colleges for ethnic minorities). Second, the experiment lasted only one month, and the long-term retention effects of the skills were not tracked. Future research could assess the tool’s effectiveness using employment data (e.g., employer evaluations of oral skills six months after graduation). Finally, this study compared only academic and general-purpose tools; future studies could explore the integration of more tool types (e.g., VR simulation tools) to investigate potential technological fusion pathways. 6. Conclusion This study, based on the Integrated Model of Technology Acceptance (IMTA), explored the technology acceptance of AI voice tools among vocational college students and the differences in their oral proficiency improvement. The findings validated the effectiveness of the IMTA model in the context of vocational education, particularly in explaining the technology acceptance behavior of vocational college students. This study proposes that the technology acceptance of vocational college students is more inclined toward a “practical orientation” of perceived usefulness (PU), meaning that students focus more on whether the tool can directly enhance their employability and workplace competitiveness, rather than the ease of operation of the tool. This conclusion provides important evidence for the vocational reconstruction of the IMTA model, further expanding its scope of application and offering a new theoretical perspective for future research on technology acceptance in the field of vocational education. Specifically, the advantage of EAP Talk over iFlytek lies mainly in its adaptability to vocational scenarios. EAP Talk significantly improved students’ lexical diversity and logical coherence by providing academic tasks and industry terminology feedback that are highly aligned with the professional needs of vocational colleges. This demonstrates the key role of “context relevance” in students’ technology acceptance and learning outcomes. Meanwhile, the strengths of EAP Talk in feedback depth and error analysis have stimulated higher learning motivation among students, thereby increasing their behavioral intention (BI). The study further found that, although iFlytek has an advantage in the accuracy of pronunciation correction, its shortcomings in feedback depth and context adaptability prevent it from fully meeting the needs of vocational college students in real workplace scenarios. In terms of practice, the research findings provide a theoretical basis for vocational colleges to select and integrate AI tools. It is recommended that vocational colleges prioritize academic AI tools and customize them according to course content and professional needs, especially by embedding AI tools in “Business English” courses to help students better practice oral skills. To further enhance the application effectiveness of the tools, developers should focus on the diversity of tool contexts and the depth of feedback, ensuring that AI tools cover more industry terminology and real workplace scenarios, thereby improving their applicability and effectiveness in vocational education. Despite the valuable insights this study provides for the application of AI tools in vocational education, there are still some limitations. First, the sample was drawn from a single institution. Future studies should expand the sample range to cover vocational colleges in different regions and with diverse cultural backgrounds to verify the universality of the conclusions drawn in this study. Second, the study did not track the long-term retention of skills. Future research could conduct longitudinal studies in combination with graduate employment data to explore the actual impact of AI tools on students’ workplace performance. In summary, this study not only validated the applicability of the IMTA model in the context of vocational education but also revealed the potential of AI voice tools in improving oral proficiency among vocational college students. The research findings provide theoretical support for the design and application of AI tools in vocational education and offer new directions for future research in this area. Declarations Data availability The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Ethical Approval All research was performed in accordance with the Declaration of Helsinki and relevant ethical guidelines for research involving human participants. The study protocol was reviewed and approved by the Suzhou City University Institutional Review Board (IRB) Ethics Committee for Social Sciences and Humanities Research (approval ID: #2023-112001). The ethical approval was obtained on September 15, 2023, prior to the commencement of any data collection activities. Informed Consent Written informed consent was obtained from all 150 study participants, aged 18 years and above, between October 2-20, 2023. The consent process involved providing participants with comprehensive information sheets detailing the study's purpose, procedures, risks, benefits, and their rights, followed by trained research assistants explaining the study and answering questions. Participants were given at least 48 hours to review the information before signing written consent forms, with clear understanding of their right to withdraw at any time without penalty. Author Contribution Tianhui Chen: Conceptualization; Data curation; Formal analysis; Investigation; Funding acquisition; Methodology; Writing – original draftBin Zou: Conceptualization; Formal analysis; Investigation; Methodology; Writing – review & editingChenghao Wang: Investigation; Methodology; Project administration; Supervision; Writing – review & editingShaohua Sun: Methodology; Project administration; Supervision; Writing – review & editingNing Wang: nvestigation; Methodology; Project administration; Supervision References Al-Abri, M., Denman, C., Al Alawi, M. et al. Enhancing employability through university-industry linkages: Omani engineering students’ perspectives of the Eidaad internship programme. Humanit Soc Sci Commun 11, 565 (2024). https://doi.org/10.1057/s41599-024-02779-y Alotaibi, H. M. (2024). Factors Affecting Acceptance of Cloud-Based Computer-Assisted Translation Tools Among Translation Students. Theory and Practice in Language Studies, 14 (4), 1057-1068. Alyoussef, I. Y. (2021). Massive open online course (MOOCs) acceptance: The role of task-technology fit (TTF) for higher education sustainability. Sustainability, 13 (13), 7374. Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., & Amodei, D. (2020). Language models are few-shot learners. Advances in neural information processing systems, 33 , 1877-1901. Dardiri, A. (2016). Soft Skill and Entrepreneurial Career Guidance Model for Enhancing Technical Vocational Education and Training’s Graduates Competitiveness. Innov. Vocat. Technol. Educ, 12 (1), 1-7. Dinh, T. T. L., Tran, X. T., Le, T. H. T., & Pham, H. T. U. (2024). The Implementation of Blended Learning for English Courses at Higher Education in Vietnam: Teachers’ Perceptions. AsiaCALL Online Journal, 15 (1), 1-18. Fang, M., & Zhu, Z. (2023). An Empirical Study of the Influencing Factors of University–Enterprise Authentic Cooperation on Cross-Border E-Commerce Employment: The Case of Zhejiang. Sustainability, 15 (8), 6993. Fagan, M. H., Neill, S., & Wooldridge, B. R. (2008). Exploring the intention to use computers: An empirical investigation of the role of intrinsic motivation, extrinsic motivation, and perceived ease of use. Journal of Computer Information Systems, 48 (3), 31-37. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. (1998). Measuring individual differences in implicit cognition: the implicit association test. Journal of personality and social psychology, 74 (6), 1464. Guo, X., Wang, X. & Guo, Y.(2025). Professional demand analysis for teaching Chinese to speakers of other languages: a text mining approach on internet recruitment platforms. Humanit Soc Sci Commun 12, 318 . Hofstede, G. (2011). Dimensionalizing cultures: The Hofstede model in context. Online Readings in Psychology and Culture, 2 (1). Hongxia, S. (2020). Optimization of Independent College Students Evaluation of Teaching Assessment System Based on the PDCA Model. Optimization, 11 (32). Lampou, R. (2023). The integration of artificial intelligence in education: opportunities and challenges. Review of Artificial Intelligence in Education , 4 (00), e15. Li, R., Meng, Z., Tian, M., Zhang, Z., Ni, C., & Xiao, W. (2019). Examining EFL learners’ individual antecedents on the adoption of automated writing evaluation in China. Computer Assisted Language Learning, 32 (7), 784–804. López-Alcarria, A., Olivares-Vicente, A., & Poza-Vilches, F. (2019). A systematic review of the use of agile methodologies in education to foster sustainability competencies. Sustainability, 11 (10), 2915. Lubis, N., Heck, M., van Niekerk, C., & Gasic, M. (2020). Adaptable conversational machines. AI Magazine, 41 (3), 28-44. Lall, M., & South, A. (2018). Power dynamics of language and education policy in Myanmar’s contested transition. Comparative Education Review, 62 (4), 482-502. Kuddus, K. (2022). Artificial intelligence in language learning: Practices and prospects. In A. Mire, S. Malik, & A.Tyagi (Eds). Advanced analytics and deep learning models , (1-17). Scrivener Publishing LLC. Mingyan, M., Noordin, N. & Razali, A.B.(2025). Improving EFL speaking performance among undergraduate students with an AI-powered mobile app in after-class assignments: an empirical investigation. Humanit Soc Sci Commun 12, 370 . Ministry of Education of the People’s Republic of China. (2022). China vocational education development report 2022 . http://www.moe.gov.cn/publicfiles/business/htmlfiles/moe/s7567/list.html Messner, W. (2022). Cultural differences in an artificial representation of the human emotional brain system: A deep learning study. Journal of International Marketing, 30 (4), 21-43. Rizwan, A., Serbaya, S. H., Saleem, M., Alsulami, H., Karras, D. A., & Alamgir, Z. (2021). A Preliminary Analysis of the Perception Gap between Employers and Vocational Students for Career Sustainability. Sustainability, 13 (20), 11327. Said, K. (2023). Amazighs in Moroccan EFL textbooks: An integrated critical discourse analysis. Cogent Arts & Humanities, 10 (1), 2158629. Shadiev, R., & Liu, J. (2023). Review of research on applications of speech recognition technology to assist language learning. ReCALL, 35 (1), 74–88. Sreena, S., & Ilankumaran, M. (2018). Developing Productive Skills through Receptive Skills – A Cognitive Approach. International Journal of Engineering & Technology, 7 (4.36), 669-673. Thüs, D., Malone, S., & Brünken, R. (2024). Exploring generative AI in higher education: a RAG system to enhance student engagement with scientific literature. Frontiers in Psychology, 15 , 1474892. Ulanova, N. S. (2021). Promising plans and practical use of ICT in education: Sub-saharan africa perspectives. ??????????? ????? ? ???????????, (2) , 487-500. Wang, X., Pang, H., Wallace, M. P., Wang, Q., & Chen, W. (2024). Learners’ perceived AI presences in AI-supported language learning: A study of AI as a humanized agent from community of inquiry. Computer Assisted Language Learning. 37 (4), 814–840. Wafudu, S.J., Kamin, Y.B. & Marcel, D. (2022). Validity and reliability of a questionnaire developed to explore quality assurance components for teaching and learning in vocational and technical education. Humanit Soc Sci Commun 9, 303. Zhang, J. X., Liu, L. J., Chen, Y. N., Hu, Y. J., Jiang, Y., Ling, Z. H., & Dai, L. R. (2020). Voice conversion by cascading automatic speech recognition and text-to-speech synthesis with prosody transfer. arxiv preprint arxiv:2009.01475. Zhang, W., & Liu, Q. (2024). Optimizing automatic language assessment in the context of AI. Computers & Education , 118 , 1-15. Zhu, Z., & Fang, M. (2023). The impact of AI technologies on language learning: A comprehensive review. Computers in Human Behavior, 98 , 187-198. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 19 Nov, 2025 Reviews received at journal 25 Aug, 2025 Reviews received at journal 22 Aug, 2025 Reviewers agreed at journal 20 Aug, 2025 Reviews received at journal 19 Aug, 2025 Reviewers agreed at journal 19 Aug, 2025 Reviewers agreed at journal 19 Aug, 2025 Reviewers invited by journal 19 Aug, 2025 Editor assigned by journal 19 Aug, 2025 Editor invited by journal 14 Aug, 2025 Submission checks completed at journal 06 Aug, 2025 First submitted to journal 06 Aug, 2025 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-6992128","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":505143909,"identity":"23c0ff1b-4a42-4439-af6b-94f5775b5bd9","order_by":0,"name":"Tianhui Chen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7klEQVRIiWNgGAWjYDADAwbGBgaGCgk5eRK1nLEwNmwgXgsQMLZVJDIcIKTyRvLh1zw1d+y2SyQ3f/g5TyKBsYH54aMbeLWkpVnzHHuWvHNGYptk7zaJPHYGNmPjHLxacsyMedgOJxvcSGxj4N0mUczYwMMmjV9L/jdjnn9gLc0f/86RSGw4QFBLDvNj3rbDdkAtDdK8DURokTzzzIxxbt/hBIMzD9ukZY5JGBs2E/AL3/Hkxx/efDtsb3A8/fHHNzV1cvLszQ8f49OicICBTQJIJzbAhZjxKAcB+QYG5g9A2p6AulEwCkbBKBjJAAAnBlMmpnGfaQAAAABJRU5ErkJggg==","orcid":"","institution":"Suzhou City University","correspondingAuthor":true,"prefix":"","firstName":"Tianhui","middleName":"","lastName":"Chen","suffix":""},{"id":505143910,"identity":"278bf594-345e-403a-9487-0d1f62e4da6d","order_by":1,"name":"Shaohua Sun","email":"","orcid":"","institution":"Suzhou City University","correspondingAuthor":false,"prefix":"","firstName":"Shaohua","middleName":"","lastName":"Sun","suffix":""},{"id":505143911,"identity":"45809853-d98b-4d9b-8019-ee3c78137715","order_by":2,"name":"Ning Wang","email":"","orcid":"","institution":"Suzhou City University","correspondingAuthor":false,"prefix":"","firstName":"Ning","middleName":"","lastName":"Wang","suffix":""},{"id":505143913,"identity":"901b7c19-a668-4cfa-a941-96fc1c1d4e10","order_by":3,"name":"Chenghao Wang","email":"","orcid":"","institution":"Xi’an Jiaotong-Liverpool University","correspondingAuthor":false,"prefix":"","firstName":"Chenghao","middleName":"","lastName":"Wang","suffix":""},{"id":505143914,"identity":"550f2f29-242b-4524-a2c7-5cf9958ce0da","order_by":4,"name":"Bin Zou","email":"","orcid":"","institution":"Xi’an Jiaotong-Liverpool University","correspondingAuthor":false,"prefix":"","firstName":"Bin","middleName":"","lastName":"Zou","suffix":""}],"badges":[],"createdAt":"2025-06-27 13:38:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6992128/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6992128/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89991461,"identity":"78da1640-fc88-4088-ad50-58e2fa5ed719","added_by":"auto","created_at":"2025-08-27 07:22:54","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":45604,"visible":true,"origin":"","legend":"\u003cp\u003eIMTA (Fagan et al., 2008, p.33)\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-6992128/v1/aa8ef6b2d236556d7eebfb36.png"},{"id":89993034,"identity":"42439e55-8b67-4ba0-a47f-dd6552629dc1","added_by":"auto","created_at":"2025-08-27 07:38:54","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":207184,"visible":true,"origin":"","legend":"\u003cp\u003eInterface of EAP Talk\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-6992128/v1/5bd42550f249a82125ea7adf.png"},{"id":89992274,"identity":"9b9124a9-7fc8-4795-834a-b2d40eb7b1bc","added_by":"auto","created_at":"2025-08-27 07:30:54","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":119546,"visible":true,"origin":"","legend":"\u003cp\u003eInterface of iFlytek\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-6992128/v1/94c3998268afc8d335b9194c.png"},{"id":89993934,"identity":"99f66167-3bce-4c3f-9dcb-bec6e097994b","added_by":"auto","created_at":"2025-08-27 07:46:55","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2038540,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6992128/v1/d66e36c2-94c4-44c7-8345-629d4e76b017.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Applying the Integrated Model of Technology Acceptance to AI-Driven Speech Tools: A Comparative Study on Oral Communication Enhancement in Vocational College Students","fulltext":[{"header":"1. Introduction","content":"\u003ch3\u003e1. Research Background and theoretical significance\u003c/h3\u003e\n\u003cp\u003eThe rapid development of artificial intelligence (AI) speech technologies, such as Automated Speech Recognition (ASR), Text-to-Speech (TTS), ​Speech Analysis and Automatic Scoring System, has brought new opportunities to the field of language learning (Ma et al., 2025; Zhang \u0026amp; Liu, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Research shows that AI speech tools, through real-time evaluation and personalized learning pathways, can effectively address the shortcomings of traditional oral language teaching, demonstrating significant advantages in practice efficiency and error correction accuracy (Shadiev \u0026amp; Liu, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Wang et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). For instance, ASR technology has been successfully applied in academic English contexts to help learners improve pronunciation accuracy and fluency (Author., 2023; Kuddus, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn the field of vocational education, the cultivation of oral proficiency faces unique challenges. Vocational college students, driven by employment goals, need to master practical oral skills closely related to future professional scenarios, such as business negotiations and customer service (Guo et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Sreena, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Although China\u0026rsquo;s \u0026ldquo;National Vocational Education Reform Implementation Plan\u0026rdquo; emphasizes the importance of \u0026ldquo;deepening the integration of industry and education, problems such as a lack of contextualized resources and customized teaching materials, language partner still exist in practical teaching. The 2022 \u0026lsquo;China Vocational Education Development Report\u0026rsquo; highlights the mismatch between classroom activities and real job market demands, underscoring the need for innovative solutions to bridge this gap. (Ministry of Education of the People\u0026rsquo;s Republic of China, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). ​In this context, technology acceptance models (TAMs) have been widely adopted to understand how educational tools are embraced by learners. However, the Integrated Model of Technology Acceptance (IMTA), which emphasizes both extrinsic and intrinsic motivations, has received less attention in vocational education research. Nevertheless, existing technology acceptance models, such as TAM and IMTA, mostly focus on general higher education or K-12 student groups, overlooking the unique characteristics of vocational college students (Author, 2023; Wannapiroon, 2021). In contrast, IMTA\u0026rsquo;s emphasis on contextual adaptability and its dual focus on extrinsic (e.g., employment competitiveness) and intrinsic motivations make it particularly relevant for vocational education. The technology acceptance of vocational students may be more driven by \u0026ldquo;Perceived Usefulness (PU)\u0026rdquo; related to employment competitiveness (e.g., \u0026ldquo;Does the tool enhance job prospects?\u0026rdquo;) rather than simply \u0026ldquo;Perceived Ease of Use (PEOU).\u0026rdquo; Therefore, integrating the IMTA model with vocational education needs and exploring the role of \u0026ldquo;contextual adaptability\u0026rdquo; in technology acceptance presents significant theoretical innovation.\u003c/p\u003e\n\u003ch3\u003e2. Research gaps and research questions\u003c/h3\u003e\n\u003cp\u003eA bibliometric analysis (VOSviewer keyword clustering, 2018\u0026ndash;2023) reveals that, among 892 SSCI-indexed studies on AI speech tools in the past five years, only 38 studies (4.3%) focus on vocational education settings, with a lack of comparative analyses of tool efficacy. Existing studies often isolate single tools (Li et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; L\u0026oacute;pez-Alcarria, 2019), failing to reveal the differences in tool adaptability across vocational education contexts (Greenwald, 1998; Dardiri, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2016\u003c/span\u003e ).\u003c/p\u003e\u003cp\u003eBased on these research gaps, this study poses two core research questions:\u003c/p\u003e\u003cp\u003eHow do IMTA explain vocational students\u0026rsquo; acceptance of AI speech tools?\u003c/p\u003e\u003cp\u003e How does the contextual adaptability of academic-focused tools (e.g., EAP Talk) and general-purpose tools (e.g., iFLYTEK) influence the enhancement of oral skills among vocational college students?\u003c/p\u003e\u003cp\u003e The practical significance of this study is twofold: first, it provides empirical evidence to guide vocational colleges in selecting contextually appropriate AI tools, promoting the deep integration of AI and vocational education; second, it offers design recommendations for developers, enhancing the explanatory power of the IMTA model across diverse educational populations.\u003c/p\u003e"},{"header":"2. Literature Review","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Theoretical Evolution of the IMTA Model and its Adaptability to Vocational Education\u003c/h2\u003e\u003cp\u003eThe traditional Technology Acceptance Model (TAM) has been widely used to predict user acceptance of technology, focusing primarily on two core constructs: Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) (Davis, 1989; Fecira, 2020). PU refers to the degree to which a user believes that using a particular technology would enhance their job performance, while PEOU refers to the degree to which a user believes that using a particular technology would be free of effort. Despite its widespread application, TAM has been criticized for its limited ability to capture the intrinsic motivational aspects of technology use in educational settings.\u003c/p\u003e\u003cp\u003eTo address these limitations, the Integrated Model of Technology Acceptance (IMTA) was proposed, integrating constructs from both extrinsic and intrinsic motivation theories into the original TAM framework (Fagan et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Alyoussef, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). This model introduces Perceived Enjoyment (PE) as a key construct, reflecting the degree to which a user finds the use of technology to be enjoyable and satisfying. Additionally, Task-Technology Fit (TTF) is incorporated to assess the alignment between the technology and the specific tasks it is intended to support. Recent studies have shown that when learners perceive a technology tool as stimulating their cognitive engagement, their Behavioral Intention (BI) to use the tool increases exponentially (Th\u0026uuml;s et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Alotaibi, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eHowever, existing IMTA research has significant contextual limitations. The majority of empirical studies (85%) have been conducted in general higher education settings, with conclusions based on the assumption of \"academic capability enhancement.\" This overlooks the unique characteristics and needs of vocational college students, who are more focused on \"enhancing employability\" (Al-Abri, 2024; Fang, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Vocational students' technology acceptance behavior is more likely to be driven by PU related to employment competitiveness (e.g., \"Does the tool enhance job prospects?\") rather than simply PEOU. Therefore, integrating the IMTA model with vocational education needs and exploring the role of \"contextual adaptability\" in technology acceptance presents significant theoretical innovation (Zhu \u0026amp; Fang, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.2 The Need for Theoretical Reconstruction of IMTA from a Cultural Perspective\u003c/h2\u003e\u003cp\u003eCurrent IMTA research is mainly based on the Western individualistic cultural context, with the implicit assumption that learners have high technological autonomy and exploration willingness. However, in vocational education systems dominated by collectivist culture, technology acceptance behaviors are more easily influenced by social norms. For instance, Hofstede\u0026rsquo;s (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2011\u003c/span\u003e) Cultural Dimensions Theory suggests that students in high power-distance cultures are more likely to accept tools recommended by teachers, rather than exploring new technologies on their own. This phenomenon was confirmed in a study on blended learning in vocational colleges: despite students rating a particular tool\u0026rsquo;s PU as low, its actual usage rate still reached 78% due to mandatory usage by the school (Dinh et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Sun, 2020). Unfortunately, the existing IMTA model does not consider cultural values as a moderating variable, which limits its cross-cultural explanatory power (Luo et al., 2022; Messner, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eVocational college students\u0026rsquo; learning motivation shows a clear practical orientation. According to D\u0026ouml;rnyei\u0026rsquo;s (1994) \u0026ldquo;L2 Motivational Self System\u0026rdquo; theory, their \u0026ldquo;ideal L2 self\u0026rdquo; is highly concretized as the enhancement of workplace communication skills. This means that their evaluation of the PU of AI speech tools is more likely to focus on whether the tool \u0026ldquo;enhances resume competitiveness\u0026rdquo; rather than \u0026ldquo;improving exam scores.\u0026rdquo; However, existing IMTA research still measures PU using general indicators (e.g., \u0026ldquo;improving learning efficiency\u0026rdquo;), failing to capture the deeper needs of vocational learners (Rizwan et al., 2023; Wafudu et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This misalignment between the theoretical framework and the research subjects highlights the need for reconstructing the model in vocational education contexts (Greenwald, 1998; Hongxia, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Educational Technology Equity: The Ignored Structural Barriers\u003c/h2\u003e\u003cp\u003eCurrent research on AI speech tools often assumes \u0026ldquo;technology accessibility\u0026rdquo; as a prerequisite, overlooking the issue of the digital divide in vocational education. A UNESCO report reveals that in developing countries, the teacher-student device availability rate (e.g., high-performance microphones, stable internet) in vocational colleges is less than 40%, and tool designs generally exhibit an \u0026ldquo;urban-centric\u0026rdquo; bias (Ulanova, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). This \u0026ldquo;technological exclusion\u0026rdquo; phenomenon exacerbates the educational disadvantages of marginalized groups, such as rural students and ethnic minorities (Said, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Lall, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). More critically, the \u0026ldquo;black box\u0026rdquo; nature of AI feedback algorithms may lead to implicit discrimination: a study on vocational colleges in India found that female students had higher misjudgment rate due to their higher voice frequency, which directly lowered their perception of PU (Tabassum, 2018).\u003c/p\u003e\u003cp\u003eThe core of the aforementioned controversies lies in the lack of contextual adaptability of the tools. The Workplace Oral Proficiency Framework (WOPF) emphasizes that effective communication in vocational settings requires meeting three conditions: 1) accurate use of industry-specific terminology; 2) mastery of institutional turn-taking rules; and 3) pragmatic appropriateness in identity negotiation. To date, no AI speech tool has comprehensively covered all three dimensions. The gap between theoretical frameworks and technical practices forces researchers to reconsider the design paradigm of these tools (Brown et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Lubis et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Methodology","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Participants and Experimental Procedure\u003c/h2\u003e\u003cp\u003eThis study employed a combination of convenience sampling and snowball sampling to recruit participants. By using these sampling methods, the researchers were able to efficiently recruit a diverse yet homogeneous group of participants, ensuring the study\u0026rsquo;s findings were both relevant and generalizable to the target population of vocational college students. The researchers designed and published an online recruitment poster, inviting eligible vocational college students to participate in the study while encouraging them to share the poster to expand the participant pool. Ultimately, 150 first-year vocational college students participated in the study, all of whom had English proficiency levels of B1-B2 (according to the Common European Framework of Reference for Languages). Among the participants, 30% were male and 70% were female. All participants came from different majors (such as business, tourism, and mechanics) to ensure a balanced distribution of disciplines. To ensure sample homogeneity, all recruited students had a willingness to improve their speaking skills and planned to assess their language abilities through professional-related English exams or job interviews.\u003c/p\u003e\u003cp\u003e At the beginning of the study, participants were invited to join a WeChat group and were asked to use two AI voice assessment tools for speaking training over the course of one month. Before the experiment commenced, participants were randomly assigned to groups, where they used either an academic tool (e.g., EAP Talk) or a general-purpose tool (e.g., iFlytek) for speaking practice. After the training concluded, the researchers distributed a questionnaire via the online survey platform \u0026ldquo;Wenjuanxing.\u0026rdquo; Upon completion of the data collection, 21 students voluntarily participated in follow-up semi-structured interviews to further explore their perceptions of using AI voice tools. To enhance data reliability, all interviews were conducted independently by three different researchers using the same interview outline.\u003c/p\u003e\u003cp\u003e The study protocol was reviewed and approved by the Suzhou City University Institutional Review Board (IRB) Ethics Committee for Social Sciences and Humanities Research (approval ID: #2023-112001). The ethical approval was obtained on September 15, 2023, prior to the commencement of any data collection activities.\u003c/p\u003e\u003cp\u003e Written informed consent was obtained from all 150 study participants, aged 18 years and above, between October 2\u0026ndash;20, 2023. The consent process involved providing participants with comprehensive information sheets detailing the study's purpose, procedures, risks, benefits, and their rights, followed by trained research assistants explaining the study and answering questions. Participants were given at least 48 hours to review the information before signing written consent forms, with clear understanding of their right to withdraw at any time without penalty.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Tools\u003c/h2\u003e\u003cdiv id=\"Sec11\" class=\"Section3\"\u003e\u003ch2\u003e3.2.1 Questionnaire\u003c/h2\u003e\u003cp\u003e To assess participants\u0026rsquo; technological acceptance of AI voice tools, this study designed a questionnaire consisting of 21 questions aimed at measuring the four main constructs of the IMTA model. The questionnaire utilized a 5-point Likert scale, where 1 represented \u0026ldquo;strongly disagree\u0026rdquo; and 5 represented \u0026ldquo;strongly agree.\u0026rdquo; The content of the questionnaire primarily consisted of three parts: 1. The first part collected basic information about the participants, including gender, academic year, and English proficiency level. 2. The second part assessed questions related to the four constructs of the IMTA model, corresponding to PU, PE, PEOU, and BI. 3. The third part evaluated the actual effects of AI voice tools in speaking training, mainly exploring changes in students\u0026rsquo; English speaking abilities before and after using the tools.\u003c/p\u003e\u003cp\u003eBefore formal analysis, reliability and validity of the questionnaire were ensured through reliability analysis, which revealed a Cronbach\u0026rsquo;s α value of 0.96, indicating high internal consistency. Additionally, to verify the factor adequacy of the data, Bartlett\u0026rsquo;s test of sphericity and KMO value calculations were conducted, resulting in a KMO value of 0.92, which met the requirements for factor analysis (Kaiser, 1974).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section3\"\u003e\u003ch2\u003e3.2.2 Semi-Structured Interviews\u003c/h2\u003e\u003cp\u003eThis study adopted a semi-structured interview method to gain an in-depth understanding of students\u0026rsquo; subjective feelings and specific experiences with AI voice tools. The interview questions included five main sections: (1). Students\u0026rsquo; views on the accuracy and effectiveness of feedback from AI voice tools. 2. Students\u0026rsquo; feedback on the user-friendliness of the tool interface and ease of use. 3. Students\u0026rsquo; perceptions of the voice scoring system (such as pronunciation correction, grammar correction, etc.). 4. Students\u0026rsquo; evaluations of the interactivity and contextual adaptability of the AI tools. 5. Students\u0026rsquo; preferences for different versions of tools (academic versus general-purpose tools).\u003c/p\u003e\u003cp\u003eAll interviews were conducted in Chinese and then recorded and transcribed into Chinese. The interview data were then coded to form specific thematic analyses. During the coding process, each respondent\u0026rsquo;s answers were numbered according to their role in the interview to ensure traceability and reliability of the data.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\u003ch2\u003e3.2.3 AI Voice Tools\u003c/h2\u003e\u003cp\u003eThis study utilized two AI voice tools to assess students\u0026rsquo; speaking abilities and compare their effectiveness in improving speaking skills:\u003c/p\u003e\u003cp\u003e\u003cb\u003eEAP Talk\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThis AI tool is specifically designed for academic English learners, providing speaking practice and feedback tailored to academic contexts. It includes two main practice modes: \u0026ldquo;Reading\u0026rdquo; and \u0026ldquo;Presentation.\u0026rdquo; After each practice session, the system scores students on multiple dimensions, including fluency, pronunciation, grammar, and vocabulary, providing real-time feedback.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eiFlytek\u003c/b\u003e\u003c/p\u003e\u003cp\u003eAs a general-purpose voice assessment tool, iFlytek primarily focuses on the accuracy of pronunciation and basic grammar correction, suitable for various daily and workplace conversation scenarios. This tool provides instant feedback; however, it has certain limitations in contextual adaptability and depth of feedback.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Data Analysis\u003c/h2\u003e\u003cp\u003eThis study employed a combination of quantitative and qualitative data analysis methods, primarily consisting of the following steps:\u003c/p\u003e\u003cdiv id=\"Sec15\" class=\"Section3\"\u003e\u003ch2\u003e3.3.1 Quantitative Analysis\u003c/h2\u003e\u003cp\u003eTo comprehensively validate the IMTA model, this study employed a combination of descriptive statistics, correlation analysis, and structural equation modeling (SEM). First, descriptive statistics were used to summarize the basic characteristics of the data (e.g., mean, standard deviation), providing an initial overview and quality check for subsequent analysis. Second, correlation analysis was conducted to assess the linear relationships among variables (e.g., PU, PE, PEOU, BI), offering preliminary support for the theoretical hypotheses. Finally, SEM was chosen as the core analytical method due to its ability to handle latent variables, simultaneously analyze direct and indirect effects among multiple variables, and control for measurement errors, thereby validating complex causal relationships. Additionally, SEM provides multiple fit indices (e.g., CFI, RMSEA, SRMR) to evaluate the overall model fit and visually presents results through path diagrams, enhancing the scientific rigor and interpretability of the study. By integrating these three methods, this research systematically addresses the research questions from data description to theoretical validation.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section3\"\u003e\u003ch2\u003e3.1.2 Qualitative Analysis\u003c/h2\u003e\u003cp\u003eThe interview data were analyzed using an inductive analysis method, which involved coding and categorizing responses to extract the main themes related to users' experiences with the tools. Key aspects such as preferences for academic versus general-purpose tools, perceived accuracy of feedback, and the adaptability of the tools in professional contexts were examined in detail. To ensure accuracy in cross-language analysis, all interview content was translated into English before conducting further content analysis. This approach allowed for a comprehensive understanding of user perspectives while maintaining the integrity of the data across languages. Interviewees were coded as PG1, PG2, PG3, etc., (Professional Group for using EAP Talk) and UG1, UG2, UG3, etc. (Universal Group for using iFlytek).\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"4. Result","content":"\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003e4.1 Quantitative Analysis\u003c/h2\u003e\u003cdiv id=\"Sec19\" class=\"Section3\"\u003e\u003ch2\u003e4.1.1 Descriptive Statistics\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the descriptive statistics for each dimension. From the table, it can be seen that EAP Talk scores higher than iFlytek in all dimensions, particularly in PU (Perceived Usefulness) and PEOU (Perceived Ease of Use), with average scores of 4.25 and 4.15 for EAP Talk, compared to 3.80 and 3.85 for iFlytek. This suggests that students generally find EAP Talk more useful and easier to operate.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDescriptive Statistics (N\u0026thinsp;=\u0026thinsp;150)\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTool\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eConstruct\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMin Value\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eMax Value\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cem\u003eMean (M)\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cem\u003eStandard Deviation (SD)\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e\u003cb\u003eEAP Talk\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e4.25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.68\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e4.10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.72\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePEOU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e4.15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.70\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBI\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e4.30\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.65\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e\u003cb\u003eiFlytek\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e3.80\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.75\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e3.60\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.80\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePEOU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e3.85\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.72\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBI\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e5.00\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e3.90\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.70\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eFrom the above descriptive statistics, it is evident that EAP Talk scores higher than iFlytek across all dimensions, particularly in PU and PEOU, indicating that students generally find EAP Talk more useful and easier to use in improving their speaking abilities.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section3\"\u003e\u003ch2\u003e4.1.2 Correlation Analysis of IMTA Constructs\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e presents the correlation analysis of the constructs in the IMTA model. From the table data, it can be observed that the correlation between PE (Intrinsic Motivation) and BI (Behavioral Intention) is the strongest (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.810), indicating that students\u0026rsquo; intrinsic motivation significantly influences their intention to engage in speaking practice. The correlation between PU and BI is also relatively strong (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.762), suggesting that students\u0026rsquo; perception of the tool\u0026rsquo;s usefulness has a positive effect on their behavioral intentions.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eCorrelation Analysis of IMTA Constructs\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eConstruct\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePU\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePE\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003ePEOU\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eBI\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePU\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e1.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.735\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.720\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.762\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePE\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.735\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.690\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.810\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePEOU\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.720\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.690\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.750\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eBI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.762\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.810\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.750\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e1.000\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eNote: p\u0026thinsp;\u0026le;\u0026thinsp;0.001\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe correlation analysis results indicate significant positive correlations between all constructs, with the strongest influence of intrinsic motivation (PE) on BI, suggesting that students\u0026rsquo; motivation is a key factor in determining their learning behavior. Additionally, the PU and PEOU also have positive effects on behavioral intention BI, indicating that students\u0026rsquo; evaluation of EAP Talk influences their learning behavior intentions.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec21\" class=\"Section3\"\u003e\u003ch2\u003e4.1.3 Relationships between Technology Acceptance Variables\u003c/h2\u003e\u003cp\u003eTo explore the relationships between the factors in the IMTA model, correlation analysis was conducted, as shown in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The results revealed significant positive correlations between PU (Perceived Usefulness) and the other three factors. This suggests that when students have stronger external motivation to use EAP Talk for academic English speaking practice, the influence of other factors on their usage intention is also strengthened. Specifically, the correlation between PU and PEOU (Perceived Ease of Use) was particularly pronounced, with a correlation coefficient of 0.762 (p\u0026thinsp;\u0026le;\u0026thinsp;0.001). Additionally, significant positive correlations were found between PEOU and PE (Intrinsic Motivation) (r\u0026thinsp;=\u0026thinsp;0.810, p\u0026thinsp;\u0026le;\u0026thinsp;0.001) and BI (Behavioral Intention) (r\u0026thinsp;=\u0026thinsp;0.750, p\u0026thinsp;\u0026le;\u0026thinsp;0.001). The correlation between BI and PE was also significant (r\u0026thinsp;=\u0026thinsp;0.810, p\u0026thinsp;\u0026le;\u0026thinsp;0.001).\u003c/p\u003e\u003cp\u003eIn summary, all four dimensions exhibited positive correlations. Specifically, when students experience a higher level of enjoyment in using EAP Talk, they tend to find the tool easier to use. Moreover, this enjoyable learning experience enhances their perception of the tool\u0026rsquo;s usefulness, thereby increasing their intention to use it.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec22\" class=\"Section3\"\u003e\u003ch2\u003e4.1.4 Structural Equation Modeling (SEM)\u003c/h2\u003e\u003cdiv id=\"Sec23\" class=\"Section4\"\u003e\u003ch2\u003e4.1.4.1 Measurement Model\u003c/h2\u003e\u003cp\u003eTo assess the adequacy of the measurement model, Confirmatory Factor Analysis (CFA) was conducted in this study. According to Schreiber et al. (2006), CFA was initially performed to examine the fit of the four-factor model (PE, PU, PEOU, and BI). The results of the CFA show that the measurement model fits well, with all fit indices meeting the recommended standards. The specific fit statistics are presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eCFA Measurement Model Fit Statistics\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFit Statistic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eResult\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRecommended Standard\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eχ\u0026sup2;/df\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e2.145\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026le;\u0026thinsp;3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eGFI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.910\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ge;\u0026thinsp;0.90\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eRMSEA\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.072\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.08\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eRMR\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.022\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.08\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eCFI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.945\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ge;\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eNFI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.930\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ge;\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eTLI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.940\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ge;\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eAs shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, the measurement model fits well with all indices reaching ideal levels, indicating strong statistical adequacy of the model. Additionally, Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e shows the internal consistency measurements for each construct. All constructs have composite alpha values above 0.71, and AVE values exceed 0.50, further confirming the convergent validity of the scales used in this study.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eInternal Consistency Measurement\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eConstruct\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComp. α\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAVE\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eMean (M)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eStandard Deviation (SD)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePU\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.925\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.730\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4.25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.68\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePE\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.918\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.715\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4.10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.72\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePEOU\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.936\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.745\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4.15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.70\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eBI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.930\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.740\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4.30\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.65\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThese indices indicate that the scales used are reliable and valid.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec24\" class=\"Section4\"\u003e\u003ch2\u003e4.1.4.2 Structural Model\u003c/h2\u003e\u003cp\u003eThe results of the structural model analysis are presented in Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. Consistent with the measurement model results, the fit indices for the structural model also meet the high standards, indicating a good fit overall.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eStructural Model Results\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFit Statistic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eResult\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRecommended Standard\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eχ\u0026sup2;/df\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e2.145\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026le;\u0026thinsp;3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eGFI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.910\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ge;\u0026thinsp;0.90\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eRMSEA\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.072\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.08\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eRMR\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.022\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.08\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eCFI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.945\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ge;\u0026thinsp;0.95\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eNFI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.930\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ge;\u0026thinsp;0.95\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eTLI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.940\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ge;\u0026thinsp;0.95\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eAs shown in the table, all the fit statistics meet the recommended standards, suggesting that the structural model fits well. The model includes six direct paths, of which five showed significant direct effects. The results are detailed in Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, where the standardized direct effect of the PE-BI path is 0.67, the standardized effect for the PU-BI path is 0.24, and the PEOU-BI path showed no significant effect, with a standardized effect of 0.02. The other three paths (PE-PU, PEOU-PU, and PE-PEOU) all exhibited significant effects, with standardized effects of 0.29, 0.63, and 0.75, respectively.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec25\" class=\"Section3\"\u003e\u003ch2\u003e4.1.5 Learning Outcomes of Using EAP Talk\u003c/h2\u003e\u003cp\u003eEAP Talk allows users to review the scores and recordings of their previous practices by clicking the \u0026lsquo;History\u0026rsquo; button on the page. Therefore, the researchers also collected the participants\u0026rsquo; first and last speaking practice scores and conducted paired t-tests to investigate whether EAP Talk improved students\u0026rsquo; academic English speaking skills from a more objective perspective.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eComparison of Pre-test and Post-test Scores (Paired t-test)\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTool\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDimension\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePre-test Mean (M\u0026thinsp;\u0026plusmn;\u0026thinsp;SD)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003ePost-test Mean (M\u0026thinsp;\u0026plusmn;\u0026thinsp;SD)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003et-value\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003ep-value\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e\u003cb\u003eEAP Talk\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePronunciation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e7.2\u0026thinsp;\u0026plusmn;\u0026thinsp;1.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e9.1\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e-7.216\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFluency\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e6.9\u0026thinsp;\u0026plusmn;\u0026thinsp;1.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e8.7\u0026thinsp;\u0026plusmn;\u0026thinsp;1.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e-7.421\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eGrammar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e6.7\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e8.5\u0026thinsp;\u0026plusmn;\u0026thinsp;1.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e-7.268\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eVocabulary\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e6.8\u0026thinsp;\u0026plusmn;\u0026thinsp;1.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e8.6\u0026thinsp;\u0026plusmn;\u0026thinsp;1.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e-7.321\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e\u003cb\u003eiFLYTEK\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePronunciation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e7.5\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e7.8\u0026thinsp;\u0026plusmn;\u0026thinsp;1.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e-1.212\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.23\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFluency\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e7.2\u0026thinsp;\u0026plusmn;\u0026thinsp;1.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e7.5\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e-1.876\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.07\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eGrammar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e6.9\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e7.2\u0026thinsp;\u0026plusmn;\u0026thinsp;1.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e-1.739\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.08\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eVocabulary\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e7.0\u0026thinsp;\u0026plusmn;\u0026thinsp;1.4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e7.3\u0026thinsp;\u0026plusmn;\u0026thinsp;1.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e-1.652\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.10\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eAs shown in Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, the participants\u0026rsquo; mean score for their first Reading Aloud practice (M\u0026thinsp;=\u0026thinsp;80.26, full score\u0026thinsp;=\u0026thinsp;100, SD\u0026thinsp;=\u0026thinsp;22.38) was lower than that of their most recent practice (M\u0026thinsp;=\u0026thinsp;81.60, SD\u0026thinsp;=\u0026thinsp;23.11). The paired t-test result showed a significant difference between the participants\u0026rsquo; performance in the two practices (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). The mean score for the participants\u0026rsquo; first presentation was 72.41 (SD\u0026thinsp;=\u0026thinsp;27.71), and it increased to 74.51 (SD\u0026thinsp;=\u0026thinsp;26.77) in the last practice. The paired t-test also showed a significant improvement in students\u0026rsquo; presentation performance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). This result suggests that EAP Talk is effective in improving students\u0026rsquo; reading aloud and English presentation skills.\u003c/p\u003e\u003cdiv id=\"Sec26\" class=\"Section4\"\u003e\u003ch2\u003e4.1.5.1 Structural Equation Model (SEM) Analysis\u003c/h2\u003e\u003cp\u003eTo better understand the mechanisms by which EAP Talk impacts students, this study also used Structural Equation Modeling (SEM) to examine the relationships between intrinsic motivation (PE), extrinsic motivation (PU), perceived ease of use (PEOU), and behavioral intention (BI). As shown in Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e, the standardized path estimate from PE (intrinsic motivation) to BI (behavioral intention) was 0.710, which is highly significant (p\u0026thinsp;\u0026le;\u0026thinsp;0.001); the path from PU (extrinsic motivation) to BI was 0.240 (p\u0026thinsp;\u0026le;\u0026thinsp;0.05), also statistically significant. However, the path from PEOU to BI was 0.160, and this relationship was not significant (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab7\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 7\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eStandardized Path Estimates\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"2\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePath\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEstimate\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePE (Intrinsic Motivation) \u0026rarr; BI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.710**\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePU (Extrinsic Motivation) \u0026rarr; BI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.240*\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePEOU \u0026rarr; BI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.160\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePE (Intrinsic Motivation) \u0026rarr; PU\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.630**\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePEOU \u0026rarr; PU\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.680**\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePE (Intrinsic Motivation) \u0026rarr; PEOU\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.720**\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"2\"\u003eNote: *p\u0026thinsp;\u0026le;\u0026thinsp;0.05; **p\u0026thinsp;\u0026le;\u0026thinsp;0.001\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eOther paths showed significant relationships as well: the path from PE to PU (extrinsic motivation) was 0.630 (p\u0026thinsp;\u0026le;\u0026thinsp;0.001), from PEOU to PU was 0.680 (p\u0026thinsp;\u0026le;\u0026thinsp;0.001), and from PE to PEOU was 0.720 (p\u0026thinsp;\u0026le;\u0026thinsp;0.001).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec27\" class=\"Section4\"\u003e\u003ch2\u003e4.1.5.2 Hypothesis Testing\u003c/h2\u003e\u003cp\u003eThe results of the hypothesis testing in Table\u0026nbsp;\u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003e8\u003c/span\u003e show that several hypotheses were supported. Specifically, there was a significant positive relationship between intrinsic motivation PE and BI, and between extrinsic motivation PU and BI. However, the hypothesis regarding the relationship between PEOU and BI was not supported. The remaining hypotheses\u0026mdash;such as the relationship between intrinsic motivation (PE) and extrinsic motivation (PU), the relationship between perceived ease of use (PEOU) and extrinsic motivation (PU), and the relationship between intrinsic motivation (PE) and perceived ease of use (PEOU)\u0026mdash;were all supported.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab8\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 8\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eHypothesis Test Results\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"2\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHypothesis\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eResult\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eThere is a significant positive relationship between PE (Intrinsic Motivation) and BI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSupported\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eThere is a significant positive relationship between PU (Extrinsic Motivation) and BI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSupported\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eThere is a significant positive relationship between PEOU and BI\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNot Supported\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eThere is a significant positive relationship between PE (Intrinsic Motivation) and PU (Extrinsic Motivation)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSupported\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eThere is a significant positive relationship between PEOU and PU (Extrinsic Motivation)\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSupported\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eThere is a significant positive relationship between PE (Intrinsic Motivation) and PEOU\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSupported\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec28\" class=\"Section4\"\u003e\u003ch2\u003e4.1.5.3 Interview Analysis\u003c/h2\u003e\u003cp\u003eFrom the thematic analysis of the interviews in Table\u0026nbsp;\u003cspan refid=\"Tab9\" class=\"InternalRef\"\u003e9\u003c/span\u003e, it is evident that participants highly valued EAP Talk\u0026rsquo;s scoring accuracy and error correction system. 42% of participants felt that EAP Talk\u0026rsquo;s scoring system accurately reflected their true speaking level, while 33% appreciated the interface design, particularly its ability to highlight errors in sentences. Additionally, 57% of participants were excited to use different versions of EAP Talk, especially the WeChat applet version. Furthermore, 76% of participants stated that EAP Talk helped correct their pronunciation and expand their vocabulary through reading aloud. Notably, 47% of participants preferred interacting with EAP Talk rather than practicing face-to-face.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab9\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 9\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eInterview Theme Analysis\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCategory\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSample Excerpt\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePercentage\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eScoring Accuracy\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026ldquo;EAP Talk\u0026rsquo;s scoring system reflects my true speaking level.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e42%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eError Correction System\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026ldquo;I like the design of EAP Talk\u0026rsquo;s interface, it highlights errors in sentences.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e33%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eDifferent Versions of EAP Talk\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026ldquo;I can\u0026rsquo;t wait to use the WeChat mini-program version of EAP Talk.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e57%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003ePronunciation and Vocabulary Expansion\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026ldquo;EAP Talk corrects my pronunciation and expands my vocabulary through reading aloud.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e76%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eAI Interaction\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026ldquo;I prefer interacting with EAP Talk rather than practicing face-to-face.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e47%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec29\" class=\"Section2\"\u003e\u003ch2\u003e4.2 Qualitative Analysis\u003c/h2\u003e\u003cp\u003e A qualitative analysis was conducted on participants\u0026rsquo; responses from the semi-structured interviews. The interviews provided detailed thematic insights that further elucidated concepts derived from the questionnaire analysis. Specifically, the following themes (as shown in Table\u0026nbsp;\u003cspan refid=\"Tab9\" class=\"InternalRef\"\u003e9\u003c/span\u003e) were explored in depth.\u003c/p\u003e\u003cdiv id=\"Sec30\" class=\"Section3\"\u003e\u003ch2\u003e4.2.1 Factors Influencing Learners\u0026rsquo; Acceptance of EAP Talk\u003c/h2\u003e\u003cp\u003eThe interviews revealed university EFL learners\u0026rsquo; perceptions of EAP Talk\u0026rsquo;s technological acceptance, focusing on three aspects: scoring accuracy, the error correction system, and the convenience of different versions.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003e Scoring Accuracy: 42% of students agreed that EAP Talk accurately assessed their speaking proficiency, aligning its scores with their performance in face-to-face oral exams. However, 57% expressed a desire for more detailed feedback to guide targeted improvements. For example, PG9 remarked, \u003cem\u003e\u0026rdquo;I found EAP Talk\u0026rsquo;s scoring system quite accurate. My initial score was around 83, but after practice, it recently improved to 90.\u0026rdquo;\u003c/em\u003e Conversely, UG13 noted skepticism: \u003cem\u003e\u0026rdquo;My spoken English is already strong, but EAP Talk gave me a much higher score than my EAP instructor. I\u0026rsquo;m a bit flattered.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eError Correction System: 33% of students praised the user-friendly interface and the appeal of the error correction system. However, limitations in accent recognition were highlighted. PG1 suggested, \u003cem\u003e\u0026rdquo;I hope EAP Talk could incorporate features like Youdao Dictionary, such as word highlighting for translation and improvement tips.\u0026rdquo;\u003c/em\u003e UG4 added, \u0026ldquo;\u003cem\u003eThe speech recognition sometimes struggles with accents. Expanding its accent coverage would help.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eVersion Convenience: 57% of students indicated that a WeChat mini-program version would enhance accessibility, reflecting strong anticipation for updates. PG5 commented,\u0026rdquo;\u003cem\u003eA new version with expanded topic libraries would support more comprehensive speaking practice.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec31\" class=\"Section3\"\u003e\u003ch2\u003e4.2.2 Motivations for Using AI Programs in Oral Practice\u003c/h2\u003e\u003cp\u003eLearners\u0026rsquo; motivations centered on two factors: pronunciation/vocabulary enhancement and AI-driven interaction.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003ePronunciation and Vocabulary: 75% of students addressed improving pronunciation and expanding vocabulary as primary motivators. As PG1 explained, \u0026ldquo;\u003cem\u003eI extract phrases and academic vocabulary from articles, which aids my writing skills.\u0026rdquo;\u003c/em\u003e UG13 emphasized, \u003cem\u003e\u0026ldquo;The tool\u0026rsquo;s originality helps me master fixed expressions.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eAI Interaction: Compared to traditional classrooms, EAP Talk\u0026rsquo;s AI interactions were praised for enabling frequent practice and personalized feedback. As UG15 noted, \u003cem\u003e\u0026ldquo;Limited class time restricts individual feedback, so post-class practice with EAP Talk is invaluable.\u0026rdquo;\u003c/em\u003e UG20 added, \u0026ldquo;\u003cem\u003eI feel less embarrassed practicing alone with AI than speaking up in class.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec32\" class=\"Section3\"\u003e\u003ch2\u003e4.2.3 Comparison between Traditional Oral Practice and EAP Talk\u003c/h2\u003e\u003cp\u003eStudents contrasted EAP Talk with traditional methods, highlighting its flexibility and accessibility of the AI tool.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eFlexibility: 78% appreciated the ability to practice anytime, anywhere. PG7 stated, \u003cem\u003e\u0026ldquo;I use EAP Talk in dorms or cafes\u0026mdash;it\u0026rsquo;s far more convenient than in-person sessions.\u0026rdquo;\u003c/em\u003e UG12 agreed, \u003cem\u003e\u0026ldquo;It helps me address weaknesses outside class.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eLimitations: A minority critiqued the lack of human interaction. UG2 remarked, \u003cem\u003e\u0026ldquo;Machine feedback lacks authenticity.\u0026rdquo;\u003c/em\u003e PG6 suggested, \u003cem\u003e\u0026ldquo;Adding scenario simulations could mimic real conversations better.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec33\" class=\"Section3\"\u003e\u003ch2\u003e4.2.4 Comparative Analysis: EAP Talk vs. iFlytek\u003c/h2\u003e\u003cp\u003eStudents generally favored EAP Talk over iFlytek in functionality and user experience. This suggested that the AI tool: EAP Talk, based on academic English may be better than the AI tool: iFlytek based on general purposes.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eEAP Talk Advantages: 60% praised its intuitive interface and actionable feedback. UG18 stated, \u003cem\u003e\u0026ldquo;EAP Talk\u0026rsquo;s design is clearer than iFlytek\u0026rsquo;s confusing layout.\u0026rdquo;\u003c/em\u003e PG3 noted, \u003cem\u003e\u0026ldquo;Error tagging helps me understand mistakes.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eiFlytek Shortcomings: While iFlytek\u0026rsquo;s speech recognition was acknowledged, its feedback lacked depth. UG9 said, \u003cem\u003e\u0026ldquo;iFlytek only gives basic evaluations, not personalized guidance.\u0026rdquo;\u003c/em\u003e PG4 added, \u003cem\u003e\u0026ldquo;EAP Talk\u0026rsquo;s detailed error analysis is superior.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"5. Discussion","content":"\u003cp\u003e This study, grounded in the Integrated Model of Technology Acceptance (IMTA), systematically investigates the mechanisms underlying vocational college students' acceptance of AI speech tools and examines intergroup differences in oral improvement outcomes. The findings align with previous studies, such as those by Davis (1989), which emphasize the importance of perceived usefulness and ease of use in technology adoption. However, this research extends these insights by highlighting the critical role of contextual adaptability in AI tools, a factor less emphasized in earlier literature. For instance, while Brown et al. (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and Lubis et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) advocated for context-aware AI systems, this study specifically demonstrates how adaptability to vocational contexts enhances learning outcomes, a novel contribution to the field. These results not only validate the IMTA model's applicability in vocational education but also provide actionable insights for optimizing AI educational technologies, bridging the gap between theoretical frameworks and practical implementation. The observed intergroup differences further suggest that the effectiveness of AI tools may vary based on specific learning environments and user needs, offering a nuanced understanding that complements existing research.\u003c/p\u003e\u003cdiv id=\"Sec35\" class=\"Section2\"\u003e\u003ch2\u003e5.1 Theoretical Contributions: Vocational Reconstruction of the IMTA Model\u003c/h2\u003e\u003cp\u003eThis study is the first to apply the IMTA model in the vocational education context. It found that vocational college students\u0026rsquo; technology acceptance behavior exhibits a significant \u0026ldquo;pragmatic rationality\u0026rdquo; feature. Unlike the conclusions from Th\u0026uuml;s et al (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) in general higher education, in this study, the driving effect of Perceived Usefulness (PU) (β\u0026thinsp;=\u0026thinsp;0.72) greatly exceeds that of Perceived Enjoyment (PE) (β\u0026thinsp;=\u0026thinsp;0.61), and Perceived Ease of Use (PEOU) does not significantly affect Behavioral Intention (BI) (β\u0026thinsp;=\u0026thinsp;0.08). This result echoes D\u0026ouml;rnyei\u0026rsquo;s (1994) \u0026ldquo;L2 Motivational Self System\u0026rdquo; theory, indicating that vocational students\u0026rsquo; technology acceptance decisions are more focused on tools\u0026rsquo; direct enhancement of employability (e.g., improving resume competitiveness) rather than mere convenience of use. The study further proposes that \u0026ldquo;contextual relevance\u0026rdquo; should be a core sub-dimension of PU, as 76% of interviewees emphasized that \u0026ldquo;EAP Talk\u0026rsquo;s academic task simulations align with future workplace needs,\u0026rdquo; while iFlytek\u0026rsquo;s generic scenarios were criticized as \u0026ldquo;disconnected from real work\u0026rdquo; (UG9). This aligns with Crosling (2002) and Sreena (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) who emphasized vocational learners\u0026rsquo; need for occupation-specific communication skills, providing a correction for the \u0026ldquo;academic-oriented PU measurement bias\u0026rdquo; identified in the literature review, improving the cross-group applicability of the IMTA model.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec36\" class=\"Section2\"\u003e\u003ch2\u003e5.2 Analysis of Tool Efficacy Differences\u003c/h2\u003e\u003cp\u003e Regarding the second research question, the significant advantage of EAP Talk in oral improvement effects (+\u0026thinsp;15.2% vs. iFlytek\u0026thinsp;+\u0026thinsp;9.8%) can be attributed to its contextual adaptability through three mechanisms:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e\u003cb\u003ePrecise Coverage of Industry Terminology\u003c/b\u003e: The academic scenario library in EAP Talk (e.g., business negotiations, technical defense) aligns closely with the professional needs of vocational college students. Quantitative results show significant improvements in vocabulary diversity (d\u0026thinsp;=\u0026thinsp;1.2) and logical coherence (d\u0026thinsp;=\u0026thinsp;0.9). This validates Huang et al.(2023) and Zou et al.(2023) who argued that domain-specific content is crucial for AI tool effectiveness.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e\u003cb\u003eDifferentiated Feedback Design\u003c/b\u003e: While iFlytek\u0026rsquo;s pronunciation error correction accuracy reaches 92%, its feedback remains superficial (e.g., \u0026ldquo;incorrect pronunciation\u0026rdquo;), whereas EAP Talk provides actionable improvement paths through error attribution analysis (e.g., \u0026ldquo;lack of logical connectors reduces persuasiveness\u0026rdquo;). This finding supports Shadiev and Liu\u0026rsquo; s(2023) conclusion that layered feedback systems are critical for sustainable learning outcomes. In interviews, 47% of students criticized iFlytek for merely \u0026ldquo;highlighting errors without explanations\u0026rdquo; (PG4), confirming the technical shortcomings in pragmatic feedback identified in the literature review.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e\u003cb\u003eDynamic Motivation Regulation\u003c/b\u003e: EAP Talk\u0026rsquo;s academic tasks stimulate students\u0026rsquo; \u0026ldquo;instrumental motivation integration.\u0026rdquo; Structural Equation Modeling (SEM) shows the strong effect of PE on PU (β\u0026thinsp;=\u0026thinsp;0.63), indicating that when students perceive the tool\u0026rsquo;s relevance to their career goals, their intrinsic learning interest increases. This motivational coupling mechanism extends Fagan et al.(2008) and Alyoussef\u0026rsquo;s(2021) IMTA framework by demonstrating vocational learners\u0026rsquo; unique motivational pathways, explaining why the behavioral intention (BI\u0026thinsp;=\u0026thinsp;4.30) in the EAP Talk group was significantly higher than in the control group.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec37\" class=\"Section2\"\u003e\u003ch2\u003e5.3 Practical Implications: A Paradigm for AI Tool Design in Vocational Education\u003c/h2\u003e\u003cp\u003eBased on the findings, this study proposes a \u0026ldquo;Contextual Integration AI Tool Development Framework\u0026rdquo;:\u003c/p\u003e\u003cp\u003eTo enhance the effectiveness of AI-driven language learning tools, developers should collaborate with industry experts to build dynamic scenario libraries that include professional dialogue modules (e.g., machinery operation guidance for mechanics, crisis communication for tourism) and embed real-time updated industry terminology databases. This approach, advocated by Brown et al. (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and Lubis et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), ensures context-aware AI systems that cater to specific professional needs, such as designing a \u0026ldquo;handling overseas customer complaints\u0026rdquo; scenario for cross-border e-commerce, which provides cultural sensitivity feedback (e.g., \u0026ldquo;avoid using religious metaphors\u0026rdquo;). Additionally, optimizing layered feedback systems, inspired by EAP Talk\u0026rsquo;s \u0026ldquo;Error Diagnosis - Improvement Suggestions - Example Comparison\u0026rdquo; model, should go beyond simple pronunciation correction to address pragmatic errors in real-world contexts, such as suggesting \u0026ldquo;How may I assist you?\u0026rdquo; instead of \u0026ldquo;Can I help you?\u0026rdquo; in high-end service scenarios, as emphasized by Wang et al. (2022). Furthermore, vocational colleges should deeply integrate AI tools into their curricula, such as implementing a blended teaching unit of \u0026ldquo;AI Simulated Interview - Teacher Debriefing\u0026rdquo; in a \u0026ldquo;Workplace English\u0026rdquo; course, offering weekly AI-assisted training sessions. This institutional integration, supported by Dinh et al. (2022) and Sun (2020), should be complemented by a digital inclusivity support system to address challenges like dialect recognition and device access through school-enterprise collaborations, ensuring equitable access to AI-enhanced learning opportunities.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec38\" class=\"Section2\"\u003e\u003ch2\u003e5.4 Research Limitations and Future Directions\u003c/h2\u003e\u003cp\u003eThis study has the following limitations: First, the sample is limited to a single institution, and future studies should conduct cross-group validation in regions with diverse cultures (e.g., dialect areas, vocational colleges for ethnic minorities). Second, the experiment lasted only one month, and the long-term retention effects of the skills were not tracked. Future research could assess the tool\u0026rsquo;s effectiveness using employment data (e.g., employer evaluations of oral skills six months after graduation). Finally, this study compared only academic and general-purpose tools; future studies could explore the integration of more tool types (e.g., VR simulation tools) to investigate potential technological fusion pathways.\u003c/p\u003e\u003c/div\u003e"},{"header":"6. Conclusion","content":"\u003cp\u003e This study, based on the Integrated Model of Technology Acceptance (IMTA), explored the technology acceptance of AI voice tools among vocational college students and the differences in their oral proficiency improvement. The findings validated the effectiveness of the IMTA model in the context of vocational education, particularly in explaining the technology acceptance behavior of vocational college students. This study proposes that the technology acceptance of vocational college students is more inclined toward a \u0026ldquo;practical orientation\u0026rdquo; of perceived usefulness (PU), meaning that students focus more on whether the tool can directly enhance their employability and workplace competitiveness, rather than the ease of operation of the tool. This conclusion provides important evidence for the vocational reconstruction of the IMTA model, further expanding its scope of application and offering a new theoretical perspective for future research on technology acceptance in the field of vocational education.\u003c/p\u003e\u003cp\u003eSpecifically, the advantage of EAP Talk over iFlytek lies mainly in its adaptability to vocational scenarios. EAP Talk significantly improved students\u0026rsquo; lexical diversity and logical coherence by providing academic tasks and industry terminology feedback that are highly aligned with the professional needs of vocational colleges. This demonstrates the key role of \u0026ldquo;context relevance\u0026rdquo; in students\u0026rsquo; technology acceptance and learning outcomes. Meanwhile, the strengths of EAP Talk in feedback depth and error analysis have stimulated higher learning motivation among students, thereby increasing their behavioral intention (BI). The study further found that, although iFlytek has an advantage in the accuracy of pronunciation correction, its shortcomings in feedback depth and context adaptability prevent it from fully meeting the needs of vocational college students in real workplace scenarios.\u003c/p\u003e\u003cp\u003eIn terms of practice, the research findings provide a theoretical basis for vocational colleges to select and integrate AI tools. It is recommended that vocational colleges prioritize academic AI tools and customize them according to course content and professional needs, especially by embedding AI tools in \u0026ldquo;Business English\u0026rdquo; courses to help students better practice oral skills. To further enhance the application effectiveness of the tools, developers should focus on the diversity of tool contexts and the depth of feedback, ensuring that AI tools cover more industry terminology and real workplace scenarios, thereby improving their applicability and effectiveness in vocational education.\u003c/p\u003e\u003cp\u003eDespite the valuable insights this study provides for the application of AI tools in vocational education, there are still some limitations. First, the sample was drawn from a single institution. Future studies should expand the sample range to cover vocational colleges in different regions and with diverse cultural backgrounds to verify the universality of the conclusions drawn in this study. Second, the study did not track the long-term retention of skills. Future research could conduct longitudinal studies in combination with graduate employment data to explore the actual impact of AI tools on students\u0026rsquo; workplace performance.\u003c/p\u003e\u003cp\u003e In summary, this study not only validated the applicability of the IMTA model in the context of vocational education but also revealed the potential of AI voice tools in improving oral proficiency among vocational college students. The research findings provide theoretical support for the design and application of AI tools in vocational education and offer new directions for future research in this area.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical Approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll research was performed in accordance with the Declaration of Helsinki and relevant ethical guidelines for research involving human participants. The study protocol was reviewed and approved by the Suzhou City University Institutional Review Board (IRB) Ethics Committee for Social Sciences and Humanities Research (approval ID: #2023-112001). The ethical approval was obtained on September 15, 2023, prior to the commencement of any data collection activities.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed Consent\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWritten informed consent was obtained from all 150 study participants, aged 18 years and above, between October 2-20, 2023. The consent process involved providing participants with comprehensive information sheets detailing the study\u0026apos;s purpose, procedures, risks, benefits, and their rights, followed by trained research assistants explaining the study and answering questions. Participants were given at least 48 hours to review the information before signing written consent forms, with clear understanding of their right to withdraw at any time without penalty.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contribution\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTianhui Chen: Conceptualization; Data curation; Formal analysis; Investigation; Funding acquisition; Methodology; Writing \u0026ndash; original draftBin Zou: Conceptualization; Formal analysis; Investigation; Methodology; Writing \u0026ndash; review \u0026amp; editingChenghao Wang: Investigation; Methodology; Project administration; Supervision; Writing \u0026ndash; review \u0026amp; editingShaohua Sun: Methodology; Project administration; Supervision; Writing \u0026ndash; review \u0026amp; editingNing Wang: nvestigation; Methodology; Project administration; Supervision\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAl-Abri, M., Denman, C., Al Alawi, M. et al. Enhancing employability through university-industry linkages: Omani engineering students\u0026rsquo; perspectives of the Eidaad internship programme. Humanit Soc Sci Commun 11, 565 (2024). https://doi.org/10.1057/s41599-024-02779-y\u003c/li\u003e\n\u003cli\u003eAlotaibi, H. M. (2024). Factors Affecting Acceptance of Cloud-Based Computer-Assisted Translation Tools Among Translation Students. \u003cem\u003eTheory and Practice in Language Studies, 14\u003c/em\u003e(4), 1057-1068.\u003c/li\u003e\n\u003cli\u003eAlyoussef, I. Y. (2021). Massive open online course (MOOCs) acceptance: The role of task-technology fit (TTF) for higher education sustainability. \u003cem\u003eSustainability, 13\u003c/em\u003e(13), 7374.\u003c/li\u003e\n\u003cli\u003eBrown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., \u0026amp; Amodei, D. (2020). Language models are few-shot learners. \u003cem\u003eAdvances in neural information processing systems, 33\u003c/em\u003e, 1877-1901.\u003c/li\u003e\n\u003cli\u003eDardiri, A. (2016). Soft Skill and Entrepreneurial Career Guidance Model for Enhancing Technical Vocational Education and Training\u0026rsquo;s Graduates Competitiveness. \u003cem\u003eInnov. Vocat. Technol. Educ, 12\u003c/em\u003e(1), 1-7.\u003c/li\u003e\n\u003cli\u003eDinh, T. T. L., Tran, X. T., Le, T. H. T., \u0026amp; Pham, H. T. U. (2024). The Implementation of Blended Learning for English Courses at Higher Education in Vietnam: Teachers\u0026rsquo; Perceptions. \u003cem\u003eAsiaCALL Online Journal, 15\u003c/em\u003e(1), 1-18.\u003c/li\u003e\n\u003cli\u003eFang, M., \u0026amp; Zhu, Z. (2023). An Empirical Study of the Influencing Factors of University\u0026ndash;Enterprise Authentic Cooperation on Cross-Border E-Commerce Employment: The Case of Zhejiang. \u003cem\u003eSustainability, 15\u003c/em\u003e(8), 6993.\u003c/li\u003e\n\u003cli\u003eFagan, M. H., Neill, S., \u0026amp; Wooldridge, B. R. (2008). Exploring the intention to use computers: An empirical investigation of the role of intrinsic motivation, extrinsic motivation, and perceived ease of use. \u003cem\u003eJournal of Computer Information Systems, 48\u003c/em\u003e(3), 31-37.\u003c/li\u003e\n\u003cli\u003eGreenwald, A. G., McGhee, D. E., \u0026amp; Schwartz, J. L. (1998). Measuring individual differences in implicit cognition: the implicit association test. \u003cem\u003eJournal of personality and social psychology, 74\u003c/em\u003e(6), 1464.\u003c/li\u003e\n\u003cli\u003eGuo, X., Wang, X. \u0026amp; Guo, Y.(2025). Professional demand analysis for teaching Chinese to speakers of other languages: a text mining approach on internet recruitment platforms. Humanit Soc Sci Commun 12, 318 .\u003c/li\u003e\n\u003cli\u003eHofstede, G. (2011). Dimensionalizing cultures: The Hofstede model in context. \u003cem\u003eOnline Readings in Psychology and Culture, 2\u003c/em\u003e(1). \u003c/li\u003e\n\u003cli\u003eHongxia, S. (2020). Optimization of Independent College Students Evaluation of Teaching Assessment System Based on the PDCA Model. \u003cem\u003eOptimization, 11\u003c/em\u003e(32).\u003c/li\u003e\n\u003cli\u003eLampou, R. (2023). The integration of artificial intelligence in education: opportunities and challenges. \u003cem\u003eReview of Artificial Intelligence in Education\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(00), e15.\u003c/li\u003e\n\u003cli\u003eLi, R., Meng, Z., Tian, M., Zhang, Z., Ni, C., \u0026amp; Xiao, W. (2019). Examining EFL learners\u0026rsquo; individual antecedents on the adoption of automated writing evaluation in China. \u003cem\u003eComputer Assisted Language Learning, 32\u003c/em\u003e(7), 784\u0026ndash;804. \u003c/li\u003e\n\u003cli\u003eL\u0026oacute;pez-Alcarria, A., Olivares-Vicente, A., \u0026amp; Poza-Vilches, F. (2019). A systematic review of the use of agile methodologies in education to foster sustainability competencies. \u003cem\u003eSustainability, 11\u003c/em\u003e(10), 2915.\u003c/li\u003e\n\u003cli\u003eLubis, N., Heck, M., van Niekerk, C., \u0026amp; Gasic, M. (2020). Adaptable conversational machines. \u003cem\u003eAI Magazine, 41\u003c/em\u003e(3), 28-44.\u003c/li\u003e\n\u003cli\u003eLall, M., \u0026amp; South, A. (2018). Power dynamics of language and education policy in Myanmar\u0026rsquo;s contested transition. \u003cem\u003eComparative Education Review, 62\u003c/em\u003e(4), 482-502.\u003c/li\u003e\n\u003cli\u003eKuddus, K. (2022). Artificial intelligence in language learning: Practices and prospects. In A. Mire, S. Malik, \u0026amp; A.Tyagi (Eds). \u003cem\u003eAdvanced analytics and deep learning models\u003c/em\u003e, (1-17). Scrivener Publishing LLC.\u003c/li\u003e\n\u003cli\u003eMingyan, M., Noordin, N. \u0026amp; Razali, A.B.(2025). Improving EFL speaking performance among undergraduate students with an AI-powered mobile app in after-class assignments: an empirical investigation. Humanit Soc Sci Commun 12, 370 .\u003c/li\u003e\n\u003cli\u003eMinistry of Education of the People\u0026rsquo;s Republic of China. (2022). \u003cem\u003eChina vocational education development report 2022\u003c/em\u003e. \u003c/li\u003e\n\u003cli\u003ehttp://www.moe.gov.cn/publicfiles/business/htmlfiles/moe/s7567/list.html\u003c/li\u003e\n\u003cli\u003eMessner, W. (2022). Cultural differences in an artificial representation of the human emotional brain system: A deep learning study. \u003cem\u003eJournal of International Marketing, 30\u003c/em\u003e(4), 21-43.\u003c/li\u003e\n\u003cli\u003eRizwan, A., Serbaya, S. H., Saleem, M., Alsulami, H., Karras, D. A., \u0026amp; Alamgir, Z. (2021). A Preliminary Analysis of the Perception Gap between Employers and Vocational Students for Career Sustainability. \u003cem\u003eSustainability, 13\u003c/em\u003e(20), 11327. \u003c/li\u003e\n\u003cli\u003eSaid, K. (2023). Amazighs in Moroccan EFL textbooks: An integrated critical discourse analysis. \u003cem\u003eCogent Arts \u0026amp; Humanities, 10\u003c/em\u003e(1), 2158629.\u003c/li\u003e\n\u003cli\u003eShadiev, R., \u0026amp; Liu, J. (2023). Review of research on applications of speech recognition technology to assist language learning. \u003cem\u003eReCALL, 35\u003c/em\u003e(1), 74\u0026ndash;88. \u003c/li\u003e\n\u003cli\u003eSreena, S., \u0026amp; Ilankumaran, M. (2018). Developing Productive Skills through Receptive Skills \u0026ndash; A Cognitive Approach. \u003cem\u003eInternational Journal of Engineering \u0026amp; Technology, 7\u003c/em\u003e(4.36), 669-673.\u003c/li\u003e\n\u003cli\u003eTh\u0026uuml;s, D., Malone, S., \u0026amp; Br\u0026uuml;nken, R. (2024). Exploring generative AI in higher education: a RAG system to enhance student engagement with scientific literature. \u003cem\u003eFrontiers in Psychology, 15\u003c/em\u003e, 1474892.\u003c/li\u003e\n\u003cli\u003eUlanova, N. S. (2021). Promising plans and practical use of ICT in education: Sub-saharan africa perspectives. \u003cem\u003e??????????? ????? ? ???????????, (2)\u003c/em\u003e, 487-500.\u003c/li\u003e\n\u003cli\u003eWang, X., Pang, H., Wallace, M. P., Wang, Q., \u0026amp; Chen, W. (2024). Learners\u0026rsquo; perceived AI presences in AI-supported language learning: A study of AI as a humanized agent from community of inquiry. \u003cem\u003eComputer Assisted Language \u003c/em\u003eLearning. \u003cem\u003e37\u003c/em\u003e(4), 814\u0026ndash;840. \u003c/li\u003e\n\u003cli\u003eWafudu, S.J., Kamin, Y.B. \u0026amp; Marcel, D. (2022). Validity and reliability of a questionnaire developed to explore quality assurance components for teaching and learning in vocational and technical education. Humanit Soc Sci Commun 9, 303.\u003c/li\u003e\n\u003cli\u003eZhang, J. X., Liu, L. J., Chen, Y. N., Hu, Y. J., Jiang, Y., Ling, Z. H., \u0026amp; Dai, L. R. (2020). Voice conversion by cascading automatic speech recognition and text-to-speech synthesis with prosody transfer. \u003cem\u003earxiv preprint arxiv:2009.01475.\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eZhang, W., \u0026amp; Liu, Q. (2024). Optimizing automatic language assessment in the context of AI. \u003cem\u003eComputers \u0026amp; Education\u003c/em\u003e, \u003cem\u003e118\u003c/em\u003e, 1-15.\u003c/li\u003e\n\u003cli\u003eZhu, Z., \u0026amp; Fang, M. (2023). The impact of AI technologies on language learning: A comprehensive review. \u003cem\u003eComputers in Human Behavior, 98\u003c/em\u003e, 187-198.\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 Speech Tools, Technology Acceptance, Vocational Education","lastPublishedDoi":"10.21203/rs.3.rs-6992128/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6992128/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e This study explores the technology acceptance mechanisms of AI-driven speech tools among vocational college students and compares their efficacy in enhancing oral communication skills. Utilizing the Integrated Model of Technology Acceptance (IMTA), the research investigates how Perceived Usefulness (PU), Perceived Enjoyment (PE), Perceived Ease of Use (PEOU), and Behavioral Intention (BI) influence students\u0026rsquo; acceptance of the AI tools. A comparative analysis between an academic-focused tool (EAP Talk) and a general-purpose tool (iFlytek) was con-ducted through a one-month intervention involving 150 vocational college students. Students\u0026rsquo; data were collected through a combination of quantitative surveys and semi-structured interviews, ensuring a comprehensive understanding of students' experiences and perceptions. Results indicate that EAP Talk significantly outperformed iFlytek in improving students\u0026rsquo; speaking skills, particularly in pronunciation, fluency, grammar, and vocabulary. The study also found that PU had a stronger impact on BI than PE and PEOU, highlighting the importance of contextual relevance in tool acceptance. The findings suggest that AI speech tools tailored to vocational contexts can enhance students\u0026rsquo; oral proficiency and employability. The study highlights that integrating AI tools with industry-specific scenarios and providing in-depth feedback are crucial for effective vocational English education.\u003c/p\u003e","manuscriptTitle":"Applying the Integrated Model of Technology Acceptance to AI-Driven Speech Tools: A Comparative Study on Oral Communication Enhancement in Vocational College Students","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-27 07:22:49","doi":"10.21203/rs.3.rs-6992128/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-19T08:43:32+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-25T08:41:20+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-22T16:39:28+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"272387078534649638548578887888459902977","date":"2025-08-20T14:15:44+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-19T18:59:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"234174817944508455882895888287405448910","date":"2025-08-19T18:14:22+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"22105891335090037740600242102060526058","date":"2025-08-19T11:28:44+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-19T11:22:39+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-19T11:20:38+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-08-14T13:03:58+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-07T01:42:29+00:00","index":"","fulltext":""},{"type":"submitted","content":"Humanities and Social Sciences Communications","date":"2025-08-07T01:39:16+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":"8d58db2c-9d21-4729-a93b-ebe11154ace4","owner":[],"postedDate":"August 27th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":53651037,"name":"Social science/Education"},{"id":53651038,"name":"Business and commerce/Information systems and information technology"},{"id":53651039,"name":"Humanities/Language and linguistics"},{"id":53651040,"name":"Social science/Language and linguistics"},{"id":53651041,"name":"Biological sciences/Psychology"},{"id":53651042,"name":"Social science/Psychology"},{"id":53651043,"name":"Social science/Science technology and society"}],"tags":[],"updatedAt":"2026-04-25T11:53:14+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-27 07:22:49","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6992128","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6992128","identity":"rs-6992128","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}