AI-empowered applications effects on EFL learners’ engagement in the classroom and academic procrastination Authors: Yi Ma, Mingyang Chen Published: 2024 ABSTRACT AbstractBackgroundIn today’s rapidly evolving educational landscape, AI-powered applications like ChatGPT, POE, and Duolingo are revolutionizing language education, offering personalized learning experiences in EFL. However, challenges such as student engagement and academic procrastination persist. This study delves into how these AI tools impact EFL learners’ engagement and procrastination tendencies, aiming to inform effective technology integration in language instruction.ObjectivesThe primary goals of this research are to assess the influence of AI-empowered applications on affective, cognitive, and behavioural engagement and academic procrastination among EFL learners. By considering both the affective and cognitive aspects of engagement, the study aims to provide insights into optimizing EFL instruction through AI-driven tools while addressing academic procrastination challenges.MethodologyA quasi-experimental research method was employed, involving ten intact classes comprising 350 students divided into two groups. Engagement and procrastination scales were administered before and after the treatment. T-tests were utilized to analyze the data, comparing pre- and post-treatment scores between the groups.FindingsThe results indicate that the experimental group, exposed to AI-empowered applications, demonstrated significantly higher levels of affective, cognitive, and behavioural engagement than the control group. Keywords: AI-empowered applications, Academic procrastination, engagement, Chinese EFL learners’ engagement. Moreover, a substantial reduction in academic procrastination was observed among students exposed to AI-empowered applications.ConclusionsThe study underscores the potential of AI-empowered applications to enhance learner engagement and mitigate academic procrastination. The findings contribute to the pedagogical discourse surrounding technology integration and advocate adopting learner-centred approaches. The incorporation of AI-empowered applications in diverse educational settings is recommended.Keywords: AI-empowered applications, Academic procrastination, Engagement, Chinese EFL learners Introduction With the pervasive integration of Artificial Intelligence (AI) into various facets of human life, its prominence in education has notably surged in recent years. AI-empowered applications, particularly in language learning, have demonstrated the potential to affect two critical aspects of EFL education: learner engagement and academic procrastination. Engagement, encompassing cognitive, emotional, and behavioural dimensions, refers to learners’ active involvement in their educational experiences and has been positively associated with academic success [1, 2]. Conversely, academic procrastination, defined as the habitual delay of essential tasks, can detract from the effectiveness of learning efforts and hinder academic achievement in EFL students [3–8]. The growth of AI in education was projected to increase by 43% between 2018 and 2022 [3], with the Horizon Report 2019 forecasting further substantial developments in AI applications for teaching and learning [4–8]. UNESCO (2019) has also identified AI as a crucial factor contributing to the escalating scientific output in this field [9]. Moreover, a variety of AI applications, incorporating advanced analytical techniques such as Machine Learning (ML), Natural Language Processing (NLP), Artificial Neural Networks (ANNs), and Affective Computing (AC), have been extensively employed within the English as a Foreign Language (EFL) context, exerting significant impacts. The role of AI-driven tools in supporting engagement and addressing procrastination became especially relevant during the disruptions of the COVID-19 pandemic, as AI technologies facilitated the widespread adoption of online EFL teaching and learning through platforms such as Zoom and Blackboard [10]. A substantial and growing body of literature indicates that AI can enhance language teaching and learning [7–11], particularly by improving the quality of online EFL learning during the COVID-19 pandemic. Despite numerous reviews on AI [12], there is a relative scarcity of reviews focusing specifically on AI’s impact on EFL learners’ engagement and academic procrastination. Therefore, this paper seeks to investigate how AI empowers EFL teaching and learning by critically reviewing research focused exclusively on the application of AI technologies in the EFL context, aiming to inspire innovative ideas in this rapidly evolving domain [12–14]. The reviewed studies cover a broad spectrum of AI applications within the EFL context, such as automated writing evaluation systems [7], chatbots for scaffolding argumentative writing [15], and virtual reality for improving oral proficiency [14]. Additionally, teachers’ adoption of AI-based applications in the EFL context has been analyzed, revealing the determinants influencing their adoption [13]. The integration of AI technologies, including Neural Machine Translation (NMT), has also been explored in foreign language teaching and learning, indicating potential implications for the Common European Framework of Reference for Languages (CEFR) [12]. Furthermore, the use of AI in diverse language learning contexts has been examined, including studies on the effects of voice chat on EFL learners’ speaking abilities [1–3, 16, 17] and the role of AI chatbots in language learning [18–23]. The application of AI extends to affective computing, where the role of reproducibility in affective computing has been investigated [18], demonstrating the multifaceted impact of AI technologies on language education. The rationale behind this research is to explore AI’s broad potential within EFL and specifically assess engagement and procrastination trends among EFL learners. Engagement, essential for academic success, requires understanding how AI-based applications like ChatGPT and Duolingo might elevate or hinder EFL learners’ involvement in language learning activities [24–30]. This research holds significance in multiple dimensions. Firstly, it addresses the specific context of EFL education, acknowledging the unique challenges and opportunities within language learning. Secondly, by focusing on engagement and procrastination, the research aims to provide actionable insights for educators and developers to enhance the effectiveness of AI applications in EFL classrooms. This work theoretically contributes by addressing the complex interaction between AI technology and learner engagement while empirically advancing our understanding of how these applications influence motivation, engagement, and overall student achievement [30–41]. This research aims to address this gap by systematically reviewing and analyzing the existing studies that exclusively explore the influence of these AI applications on EFL learners’ engagement in the classroom and their tendencies toward procrastination. In line with this gap, the following questions are raised: Do language learners’ use of AI-empowered educational applications affect their engagement in the classroom and types of engagement? Do language learners’ use of AI-empowered educational applications affect their engagement in the classroom and types of engagement? Do language learners’ use of AI-empowered educational applications affect procrastination and types of procrastination? Do language learners’ use of AI-empowered educational applications affect procrastination and types of procrastination? Review of literature Theoretical background This study draws on the Self-Determination Theory (SDT) and the Expectancy-Value Theory (EVT) to provide a robust theoretical foundation for examining EFL student engagement and procrastination within AI-mediated learning environments. SDT postulates that individuals are more likely to engage in activities when they experience autonomy, competence, and relatedness [42]. In the context of EFL, these factors are critical as AI tools can enhance students’ perceived control over their learning and provide immediate feedback that supports skill development, which is central to engagement. EVT, on the other hand, emphasizes the role of Expectancy (students’ beliefs about their abilities to succeed) and value (importance placed on the task) in motivating academic behaviours [43]. By understanding how these theories intersect within AI-driven EFL contexts, this study aims to elucidate how engagement and procrastination influence students’ perceptions of their abilities and the relevance of tasks facilitated through AI tools. Students’ engagement Student engagement, rooted in SDT, is defined as students’ willingness, need, desire, and compulsion to participate in and succeed in the learning process [44]. In online course delivery, effective pedagogical strategies are essential for creating diverse learning opportunities and fostering engagement. Engagement goes beyond merely acquiring cognitive skills and emphasizes students’ dispositions and attitudes toward classroom experiences and lifelong learning [45]. This concept involves assessing students’ interest levels, interactions within the course, and motivation to understand the subject matter [46]. Affective factors, such as attitudes, personality, motivation, effort, and self-confidence, are closely linked to student engagement [42]. Guo et al. [45] found a positive relationship between the quality of interaction within online courses and students’ grades. By evaluating student engagement and considering these affective elements, instructors can design lessons and activities encouraging active participation [46]. Students who are motivated, involved, and invested in their educational efforts are more likely to be engaged [46]. Engagement in a course encompasses more than traditional metrics; it includes mastery of learning objectives, retention, and satisfaction, offering a holistic view of the teaching-learning dynamic [46]. Measuring engagement levels enables instructors to adjust instructional practices in response to student motivation, involvement, and attitude changes. Online learning environments offer various tools for instructors to gather informal data on student participation, such as log-in data, time spent online, module views, and self-reported information through surveys, reflections, and discussions [47]. Evaluating the academic challenge of each course involves considering factors like effort, time investment, faculty and peer interaction, active learning, and enriching educational experiences [48]. Tools like the Student Course Engagement Questionnaire (SCEQ), developed by Handelsman et al. [48], allow faculty to assess different types of engagement, including skills, emotions, participation/interaction, and performance. Reviewing both informal and formal assessments provides faculty with insights into student perceptions, leading to improvements in instructional practices that enhance learning across courses and programs [49–53]. Students’ procrastination Explored within the framework of EVT, procrastination significantly threatens academic success for students, faculty, and educational support professionals. Its severity is often highlighted in first-year experience textbooks, which urge students to address it promptly [54]. Recognized as a significant obstacle to student success and a pervasive problem in college [55–57], procrastination is frequently described as the enemy of effective time management [54]. Despite these warnings, procrastination remains widespread on college campuses [58, 59]. While timely and effective work completion is crucial for academic progression, curricula and programs’ warnings against procrastination do not necessarily reduce its occurrence. This article explores recent research on student procrastination, behavioural economics, and project management to offer a nuanced understanding of procrastination in college. The findings are used to develop a framework for student intervention within learning centres to facilitate effective workflow and promote lasting learning. Procrastination is generally defined as the tendency to delay or avoid tasks or decisions, with habitual procrastination being associated with increased stress, anxiety, and lower academic performance [60, 61]. Recent studies differentiate between active and passive procrastination [62], noting that not all procrastination results in adverse educational outcomes. Active procrastination involves a strategic delay, where students choose to work under pressure and still achieve satisfactory grades [63]. In contrast, passive procrastination typically leads to negative academic consequences [64, 65]. Inefficiencies in workflow and procrastination are common on college campuses, with estimates indicating that 70–95% of students procrastinate [66]. Despite the frequent warnings in college success curricula, about 50% of students habitually procrastinate [67]. However, not all procrastination is necessarily harmful; strategic delay can lead to similar levels of academic success. Thus, students should be encouraged to develop comprehensive and strategic plans for completing their work, emphasizing efficient scheduling and task management. The critical issue is not the delay but the total time students dedicate to academic work, which often falls short of faculty expectations. Research suggests encouraging students to strategically plan and complete academic tasks is essential for achieving effective academic performance [68, 69]. Methodology Sampling The sample consisted of 350 first-year undergraduate students from Foshan University, with 175 students assigned to the control group and 175 to the experimental group. The control group included 87 females and 88 males, while the experimental group had 82 females and 93 males. Students ranged from 18 to 20, with an average age of 18.9 years across both groups. Each group was drawn from five different academic majors, though specific disciplines varied between groups, potentially influencing outcomes. Both groups were organized into five classes with 35 students per class, creating uniform class structures. The primary difference between the groups was that the experimental group received a three-hour workshop on AI-empowered educational applications, while the control group did not. This workshop exposed the experimental group to AI tools for collaborative integration into course content, facilitating analyses of AI’s potential impact on engagement and academic procrastination. Participants’ demographic information is presented in Table 1. Table 1Demographics for both the control and experimental groupsDemographicControl GroupExperimental GroupTotal Participants175 students175 studentsGender Distribution87 females, 88 males82 females, 93 malesAge Range18–20 years18–20 yearsAverage Age18.9 years18.9 yearsAcademic MajorsFive selected majors (e.g., Business, Law, etc.)Five selected majors (e.g., Engineering, Sciences, etc.)Class StructureFive classes, 35 students eachFive classes, 35 students eachAdditional TrainingNo additional training providedThree-hour workshop on AI applications Demographics for both the control and experimental groups With regard to ethical considerations, it can be stated that it was deemed ethically acceptable for the experimental group to attend additional AI-focused training, as the primary purpose of the study was to assess the impact of educational interventions on learner engagement and academic procrastination. Both groups received equivalent educational content within their respective majors, with the additional AI workshop offered as supplementary training rather than a replacement for core learning. Participants and instructors in both groups were fully informed about the study’s objectives and design, and their consent was obtained, aligning the intervention with ethical standards. To ensure equity, the control group was also offered access to the AI workshop upon completion of the study, providing them the same opportunity to benefit from the workshop’s content. Additionally, this study supported Foshan University’s interest in exploring innovative educational approaches, offering potential value for both the students and the institution. Instruments In this investigation, a quantitative instrument was the Student Engagement Scale, a self-report tool designed to evaluate students’ active participation in classroom activities. This scale assessed engagement across three dimensions: Affective, Cognitive, and Behavioral. Participants self-assessed their engagement levels on a 5-point scale, ranging from 1 to 5, with elevated scores signifying heightened engagement [70]. The reliability of the Student Engagement Scale in this study was confirmed with a Cronbach’s alpha of 0.89 for Affective engagement, 0.87 for Cognitive engagement, and 0.91 for Behavioral engagement, indicating strong internal consistency across all dimensions. The second quantitative measure was the Procrastination Assessment Scale-Students (PASS), developed by Solomon, L. J., & Rothblum [71], renowned for its robust psychometric properties. This scale gauged procrastination across six academic tasks, prompting participants to rate the frequency, perceived challenge, and inclination to reduce procrastination for each task. Total scores ranged from 12 to 60, with higher scores indicating increased self-reported academic procrastination. The reliability of the PASS in this study was found to be high, with a Cronbach’s alpha of 0.88, supporting its reliability for assessing academic procrastination. Procedure In this comprehensive research initiative, a quasi-experimental research design was implemented, utilizing convenience sampling to meticulously select ten intact classes, each representing a distinct major, from the cohort of first-year students at Foshan University, China. The ten majors were judiciously divided into two groups—the control group and the experimental group. This strategic allocation ensured a balanced representation of disciplines, with each group comprising five majors. In the subsequent phase, students in the experimental group, alongside their respective teachers, were actively engaged in a three-hour workshop designed to develop and incorporate AI-empowered educational applications. The primary goal of this workshop was to equip students with the essential knowledge and skills required for meaningful participation in integrating AI-driven tools within their academic experience. The workshop initiation involved an introductory session on AI-empowered educational applications, emphasizing the significance of collaborative decision-making in leveraging technology for learning. Participants were systematically guided through the fundamental aspects of integrating AI tools, focusing on customizable elements such as content delivery, interactive modules, and assessment methods. The subsequent segment shared practical strategies for seamlessly integrating AI-driven educational applications, addressing the intersection of technological advancements with academic requirements. This session aimed to empower students to utilize AI tools to enrich learning experiences effectively. The workshop’s core comprised interactive sessions wherein students, guided by their teachers, actively integrated AI-empowered educational applications into specific courses. This hands-on approach facilitated the immediate application of AI principles in academic settings. To ensure clarity and address any queries, a dedicated Question and Answer session allowed students to seek clarifications, pose questions, and discuss concerns related to integrating AI-driven educational applications. After the AI workshop, the experimental group applied these AI tools within their courses for one semester. During this time, students collaboratively engaged with AI-driven applications in real course tasks, consolidating the skills and knowledge from the workshop. At the end of the 16-week semester, the post-tests were administered to all intact classes in both the control and experimental groups. Both pretest and post-test questionnaires were specifically developed to measure engagement and academic procrastination in the context of AI-empowered educational applications. Data analysis Multivariate Analysis of Covariance (MANCOVA) was selected to examine group differences because it allows for the control of covariates while assessing multiple dependent variables. This approach is particularly suitable for examining complex relationships between variables, as it accounts for potential interactions and reduces Type I error rates by evaluating dependent variables in a single analysis. Descriptive statistics, including mean and standard deviation, summarized central tendencies and group variabilities. Results Research question 1 A Multivariate Analysis of Covariance (MANCOVA) was conducted to evaluate differences in students’ engagement levels (affective, cognitive, behavioural, and total) between the Control and Experimental groups after controlling for pretest scores. Descriptive statistics, including means and standard deviations for pretest and post-test scores across all engagement dimensions, are presented in Table 2. Initial scores were comparable across groups, indicating similar baseline engagement levels before the intervention. However, post-test scores suggest that the experimental group displayed higher engagement across all dimensions than the control group. Table 2Descriptive statistics of the groups’ scores on engagementVariableGroupMeanSD N PretestAffective EngagementControl3.400.70175Experimental3.420.72175Cognitive EngagementControl3.450.68175Experimental3.460.70175Behavioural EngagementControl3.480.66175Experimental3.500.68175Post-testAffective EngagementControl3.450.65175Experimental4.231.00175Cognitive EngagementControl3.460.56175Experimental4.330.42175Behavioural EngagementControl3.500.49175Experimental4.400.80175Engagement (Total)Control10.401.80175Experimental12.932.23175 Descriptive statistics of the groups’ scores on engagement The MANCOVA results, shown in Table 3, indicated a statistically significant multivariate effect of group on engagement variables, Wilks’ Lambda = 0.85, F(4, 343) = 14.56, p < .001, partial η² = 0.15, suggesting that after controlling for pretest scores, engagement levels differed significantly between the Control and Experimental groups. This finding demonstrates that the intervention substantially influenced the students’ engagement levels. Table 3Multivariate test resultsEffectTest StatisticValueFHypothesis dfError df p Partial η²GroupWilks’ Lambda0.8514.564343< 0.0010.15Pillai’s Trace0.1614.564343< 0.0010.15Hotelling’s Trace0.1914.564343< 0.0010.15Roy’s Largest Root0.1914.564343< 0.0010.15 Multivariate test results The results of the MANCOVA analysis demonstrate that the experimental group exhibited significantly higher levels of affective, cognitive, behavioural, and overall engagement than the control group, even after controlling for initial engagement levels. This suggests that the intervention effectively enhanced student engagement across multiple dimensions. Table 4 highlights the significant effects of the intervention on each dimension of engagement, as revealed by ANCOVA analyses while controlling for pretest scores. Affective engagement showed a robust increase in the Experimental group, with F(1, 348) = 155.54, p < .001, and a large effect size (partial η² = 0.99), indicating substantial improvement in students’ emotional connection to learning. Similarly, cognitive engagement exhibited the strongest impact, with F(1, 348) = 548.52, p < .001 and an exceptionally large effect size (partial η² = 0.998), reflecting enhanced critical thinking and understanding. Behavioural engagement also improved significantly, evidenced by F(1, 348) = 338.78, p < .001 and partial η² = 0.997, demonstrating increased participation and effort in academic tasks. Finally, total engagement, encompassing all three dimensions, was significantly higher in the Experimental group, with F(1, 348) = 274.31, p < .001 and partial η² = 0.996, confirming the intervention’s broad and consistent effectiveness. Together, these results underscore the intervention’s substantial and multidimensional impact on fostering student engagement. Table 4ANCOVA results for each dimension of engagementDependent VariableFdf1df2p-valuePartial η²Affective Engagement155.541348< 0.0010.99Cognitive Engagement548.521348< 0.0010.998Behavioral Engagement338.781348< 0.0010.997Total Engagement274.311348< 0.0010.996 ANCOVA results for each dimension of engagement Research question 2 MANCOVA was also used for the second research question to analyze the differences in academic task performance and procrastination between groups, controlling for pretest scores. The descriptive statistics for these variables are presented in Table 5. While pretest scores were similar, post-test scores for the Experimental group were lower across all procrastination variables, suggesting decreased procrastination. Table 5Descriptive statistics of the groups’ scores on aspects of procrastinationVariableGroupMeanSD N PretestWriting term assignmentsControl24.304.5175Experimental24.254.6175Keeping up with weekly assignmentsControl25.505.3175Experimental25.605.2175Doing administrative tasksControl27.305.35175Experimental27.205.4175Attending meetings and classesControl23.502.91175Experimental23.602.8175Performing academic tasksControl19.304.11175Experimental19.204.0175Studying for examinationsControl17.143.96175Experimental17.204.0175Procrastination TotalControl22.234.20175Experimental22.304.1175Post-testWriting term assignmentsControl24.304.5175Experimental19.254.6175Keeping up with weekly assignmentsControl25.505.3175Experimental16.504.32175Doing administrative tasksControl27.305.35175Experimental17.604.42175Attending meetings and classesControl23.502.91175Experimental14.263.35175Performing academic tasksControl19.304.11175Experimental15.305.23175Studying for examinationsControl17.143.96175Experimental12.524.58175Procrastination TotalControl22.234.20175Experimental14.604.89175 Descriptive statistics of the groups’ scores on aspects of procrastination The MANCOVA multivariate test results, shown in Table 6, reveal a significant multivariate he multivariate results in Table 5 confirm a significant effect of group on procrastination variables, Wilks’ Lambda = 0.78, F(7, 348) = 18.92, p < .001, partial η² = 0.28, indicating significant differences in procrastination between groups. Table 6Multivariate test resultsEffectTest StatisticValueFHypothesis dfError df p Partial η²GroupWilks’ Lambda0.7818.927348< 0.0010.28Pillai’s Trace0.2218.927348< 0.0010.28Hotelling’s Trace0.2918.927348< 0.0010.28Roy’s Largest Root0.2918.927348< 0.0010.28 Multivariate test results These results indicate that the experimental intervention had a meaningful impact on reducing procrastination and improving engagement, supporting the efficacy of the AI-integrated educational applications in enhancing students’ academic engagement and reducing procrastination. As shown in Table 7, the ANCOVA results provide compelling evidence of the intervention’s significant impact on lowering procrastination across all assessed dimensions. A significant reduction was observed for writing term assignments, as indicated by F(1, 348) = 214.34, p < .001, partial η² = 0.995, demonstrating that the intervention substantially improved students’ ability to complete term assignments on time. Similarly, procrastination in keeping up with weekly assignments was significantly lower in the Experimental group (F(1, 348) = 609.18, p < .001, partial η² = 0.998), underscoring the intervention’s effectiveness in promoting consistent engagement with ongoing academic tasks. Table 7ANCOVA results for procrastination dimensionsDependent VariableFdf1df2p-valuePartial η²Writing Term Assignments214.341348< 0.0010.995Keeping Up with Weekly Assignments609.181348< 0.0010.998Doing Administrative Tasks686.061348< 0.0010.999Attending Meetings and Classes1516.371348< 0.0010.999Performing Academic Tasks127.651348< 0.0010.992Studying for Examinations203.691348< 0.0010.995Total Procrastination490.381348< 0.0010.998 ANCOVA results for procrastination dimensions Furthermore, the intervention had a profound impact on reducing procrastination in doing administrative tasks, with F(1, 348) = 686.06, p < .001, partial η² = 0.999, suggesting that students in the Experimental group were better equipped to manage essential but often overlooked academic responsibilities. Even more striking, the largest effect was observed in attending meetings and classes, where F(1, 348) = 1516.37, p < .001, partial η² = 0.999 revealed the intervention’s remarkable success in encouraging regular participation in scheduled activities. In addition, procrastination related to performing academic tasks showed a significant decrease (F(1, 348) = 127.65, p < .001, partial η² = 0.992), indicating an overall improvement in students’ ability to manage their workloads efficiently. A similarly strong effect was found for studying for examinations, where the Experimental group exhibited significantly lower procrastination (F(1, 348) = 203.69, p < .001, partial η² = 0.995), highlighting the intervention’s role in fostering timely and focused study habits. Finally, the total procrastination score revealed a substantial overall improvement (F(1, 348) = 490.38, p < .001, partial η² = 0.998), confirming the intervention’s comprehensive success in mitigating procrastination behaviours across all academic domains. These findings, with exceptionally large effect sizes throughout, emphasize the intervention’s effectiveness in addressing task-specific procrastination and fostering sustained academic discipline and time management skills. Discussion This study explores the impact of AI-empowered educational applications on EFL learners’ engagement and academic performance, aligning with the increasing emphasis on AI’s transformative role in education. Previous studies highlight the positive effects of AI in enhancing learning experiences and engagement, suggesting that AI can significantly influence student engagement’s cognitive, behavioural, and affective dimensions [1, 5, 6, 16]. The theoretical framework of this study is grounded in self-determination theory (SDT), positing that students’ intrinsic and extrinsic motivations are critical to their engagement and performance [2, 3, 17]. This aligns with our research questions aimed at assessing how AI applications impact engagement and motivation and examining the specific areas of student performance and procrastination behaviours. The quasi-experimental design employed here involved first-year students from Foshan University, divided into control and experimental groups to ensure methodological rigor. This approach allowed us to systematically examine the differences in engagement outcomes between the two groups, offering a robust assessment of the effects of AI tools on academic tasks. The quantitative results (Table 1) demonstrate statistically significant improvements in the experimental group across cognitive, behavioural, and affective engagement dimensions. These findings are consistent with studies by Pikhart et al. and Pan et al., who found that structured feedback and teacher scaffolding positively influenced student engagement and well-being in online learning environments (2,4,6). Effect sizes calculated through Cohen’s d further substantiate the practical significance of AI tools, particularly in fostering cognitive and behavioural engagement. These findings underscore the potential of AI applications to serve as useful educational tools that can motivate and engage students, which aligns with recent discussions on the role of AI in educational psychology and student motivation (2,5,9). Moreover, this study supports previous research by Ahn and Lee, who found that AI applications utilizing automatic speech recognition enhanced students’ speaking engagement and perceived learning effectiveness (6,11). The qualitative findings reveal that students in the experimental group displayed higher engagement in writing assignments and weekly academic tasks. For instance, when completing term assignments, the control group significantly differed from the experimental group, highlighting the motivational benefits of AI applications in reducing procrastination and increasing focus (1,10,13). Similar patterns were observed in weekly assignment completion, where the control group differed significantly from the experimental group. This suggests that students exposed to AI applications are more consistent in meeting deadlines. This supports findings from Layali and Al-Shlowiy, who noted that e-learning tools could help mitigate procrastination behaviours among ESL/EFL learners [9, 13, 19]. Comparatively, this study builds upon findings by Haristiani, who demonstrated that AI-driven chatbots could enhance language engagement and academic task performance among EFL learners, providing evidence for AI’s positive impact on academic engagement [14]. Additionally, Chen et al. examined neural machine translation in foreign language learning, noting that such tools facilitate learning and promote engagement by offering timely feedback and interactive support, which aligns with our findings on the engagement benefits of AI applications [10, 18, 21]. Our study also sheds light on the implications of AI integration in education, particularly concerning the evolving roles of educators. The data suggest that educators may need to shift towards more facilitative roles, guiding students in leveraging AI tools effectively rather than solely delivering content [7, 11, 15]. This calls for further exploration of ethical considerations surrounding AI in education and long-term impacts on student autonomy and dependency on technology. By integrating quantitative and qualitative data, this study contributes a comprehensive perspective to the discourse on AI in education, highlighting the potential for AI-empowered applications to transform student engagement and academic performance sustainably [20, 22, 24]. Future studies should address limitations such as this research’s short-term scope, investigating AI’s longitudinal impacts on academic engagement and performance. Furthermore, while this study demonstrates the effectiveness of AI in reducing procrastination, it opens questions regarding the specific features of AI tools that are most influential in fostering motivation and self-regulated learning in EFL contexts [12, 15, 23]. Ultimately, the results underscore the transformative potential of AI-empowered applications in creating engaging and effective learning environments, providing a foundation for future research and practical applications in AI-enhanced EFL education. The findings demonstrate that AI-empowered educational applications can significantly enhance student engagement and support effective academic behaviours. The study, conducted with first-year students at Foshan University in a quasi-experimental design, indicates that AI tools meaningfully impact affective, cognitive, and behavioural engagement dimensions, reinforcing existing research on AI’s potential in educational contexts (1,3). Quantitative results show substantial differences between control and experimental groups, particularly in cognitive and behavioural engagement, highlighting the practical effectiveness of AI-driven applications [1, 5, 9]. Qualitative data deepens these findings, revealing themes such as “Student Empowerment through Negotiation,” which aligns with existing literature on the transformative potential of AI in creating supportive learning environments [1, 7]. Other themes, including “Impact on Classroom Interaction and Communication” and “Alignment with Individual Learning Styles and Preferences,” correspond with learner-centred educational theories, underscoring the role of AI in fostering individualized and interactive learning experiences [23, 31]. The theme “Perceived Impact on Academic Performance” suggests the academic benefits of AI, reinforcing the belief that technology integration enhances engagement and academic success [13, 20]. The theme “Challenges and Opportunities in Syllabus Negotiation” also underscores the importance of providing guidance and support, highlighting potential obstacles in AI integration similar to those discussed in broader educational AI research (8,12). This study contributes to the growing discourse on AI in education, addressing the need for further investigation into long-term effects, ethical considerations, and the evolving role of educators within AI-integrated classrooms. Combining quantitative and qualitative methods, the study underscores the transformative potential of AI, suggesting that it can foster meaningful engagement and enhance students’ ability to manage academic tasks effectively. These insights align with the broader literature on AI’s educational efficacy (1,15), positioning AI as a promising tool for enhancing learning experiences and supporting students’ academic success. Implications The implications of this study underscore the transformative potential of AI-empowered educational applications in enhancing student engagement, reducing procrastination, and fostering effective academic behaviours among EFL learners. Through a quasi-experimental design with first-year students at Foshan University, this research demonstrates that AI applications significantly impact cognitive, affective, and behavioural engagement, corroborating existing literature on the efficacy of AI in educational settings. The study’s findings suggest that AI tools increase student motivation and enhance task management and focus, reinforcing theories of self-determination and learner-centred approaches. Quantitative results reveal substantial improvements in the experimental group, highlighting the practical effectiveness of AI in promoting engagement, particularly in cognitive and behavioural dimensions. Qualitative themes, such as “Student Empowerment through Negotiation” and “Perceived Impact on Academic Performance,” reveal that students perceive AI as beneficial to their academic success, aligning with theories on individualized and interactive learning experiences. Furthermore, the theme “Challenges and Opportunities in Syllabus Negotiation” underscores the need for supportive instructional guidance to maximize the benefits of AI, addressing potential obstacles in technology integration. This research contributes to the growing discourse on AI’s role in education by emphasizing its potential to create engaging, effective learning environments. It also highlights implications for educators’ evolving roles in AI-integrated classrooms, calling for continued exploration of AI’s ethical considerations and long-term impacts on student autonomy. By providing insights into both the immediate effects and the promising future of AI-enhanced learning, this study lays a foundation for further research into AI’s sustained impact on academic engagement, motivation, and performance in EFL contexts. Limitations and suggestions for further studies While studies exploring the impact of AI-empowered applications like ChatGPT, PPE, and Duolingo on EFL (English as a Foreign Language) learners’ engagement and academic procrastination have shown promising results, several limitations must be addressed to strengthen the validity and applicability of findings. Firstly, many studies suffer from tiny sample sizes or lack diversity among participants, limiting the generalizability of results. Future research should strive for more extensive and diverse samples to better represent the population of EFL learners. Secondly, the duration of most studies tends to be short-term, neglecting the potential long-term effects of AI-empowered applications. Longitudinal studies are warranted to investigate sustained impacts over an extended period. Furthermore, the absence of appropriate control groups in some studies poses challenges in attributing observed changes solely to using AI-empowered applications. Large-scale randomized controlled trials (RCTs) with proper control groups would enhance the internal validity of findings. Contextual factors such as classroom environment, teacher involvement, and institutional support are often overlooked but can significantly influence the effectiveness of AI-empowered applications. Future research should conduct contextual analyses to understand these influences better. Measurement tools used to assess engagement and procrastination may only sometimes capture the nuances of these constructs accurately. Developing and validating robust measurement tools, both quantitative and qualitative, is essential for reliable assessment. Lastly, comparative studies between AI-empowered applications can provide valuable insights into their relative effectiveness. Such studies can guide educators and developers in selecting the most suitable tools for specific educational contexts. In summary, addressing these limitations and implementing the suggested approaches in future research endeavours will contribute to a more comprehensive understanding of the impact of AI-empowered applications on EFL learners’ engagement and academic procrastination, ultimately informing more effective educational practices. Acknowledgements The author would like to thank all participants. Author contributions YM designed the study. MC collected the data. YM analyzed and interpreted the data. YM drafted the manuscript. MC proofread the paper. YM agreed to be accountable and verified the submitted version. Data availability Data is provided within the manuscript or supplementary information files. Declarations Ethics approval and consent to participate The ethical approval committee of Foshan University approved this study and issued a letter indicating that it had no side effects on the participants. All experiments and methods were carried out by relevant guidelines and regulations. Informed consent was obtained from all subjects. Consent for publication Not applicable. Clinical trial number Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement Data is provided within the manuscript or supplementary information files.