Order	Time	ID	Full information
1	13:00-13:15	IT3318	Understanding the Role of AI-Interaction Positivity in College Students’ AI Literacy Author(s): Xiao Zhou, Ningyue Li, Liu Li Presenter: Xiao Zhou, Guangdong University of Science and Technology, China Abstract: As artificial intelligence (AI) technologies become increasingly embedded in higher education, developing students’ AI literacy has emerged as a critical educational goal. This study examines the relationship between AI-Interaction positivity and AI literacy among college students and explores the predictive role of positive AI interaction experiences. Using a survey-based design, data were collected from undergraduate students through a non-probability convenience sampling approach and analyzed using partial least squares structural equation modeling (PLS-SEM). The findings indicate that AI-Interaction positivity is positively and significantly associated with AI literacy and serves as a meaningful predictor of students’ AI-related competencies. The results suggest that students who experience AI interactions as engaging, manageable, and supportive tend to demonstrate higher levels of AI literacy. From an educational perspective, this study highlights the importance of considering learners’ subjective interaction experiences with AI systems when promoting AI literacy in higher education. The findings offer practical implications for the design of AI-supported learning environments that foster positive human–AI interaction and support responsible and informed AI use.
2	13:15-13:30	IT3247	An Empirical Study on the Influencing Factors of GenAI on Learning Achievement in Basic Courses of Artificial Intelligence Major Author(s): Haibin Xie, Hanyin Zheng, Yijing Huang, Guolin Wu Presenter: Haibin Xie, Guilin University of Aerospace Technology, China Abstract: Generative Artificial Intelligence (GenAI) is significantly impacting the learning achievement of emerging engineering majors in programming education. First, a “core-extension” Structural Equation Model (SEM) framework is constructed, which incorporates four progressive core literacies: Knowledge and Skills, Methods and Abilities, Thinking and Innovation, Career Competence Development and two GenAI variables. Secondly, GenAI's direct and indirect effects on learning ability are empirically examined point to the cultivation path “Knowledge→Methods→Thinking→Career literacy”. The results confirm the significance of this progressive literacy path, with Methods and Abilities acting as the critical intermediary; GenAI exerts a direct positive effect on Methods and Abilities and an indirect positive effect on Career Competence Development via literacy chain transmission. Finally, some targeted and operable instructional implications are proposed for the quality assessment and teaching optimization of GenAI-integrated programming courses in emerging engineering education.
3	13:30-13:45	IT3365	Development and Validation of the Pre-Service Preschool Teachers' AIGC Readiness Scale Author(s): Yuejiao Gu, Yichun Qian, Binen Ye, Suman Han Presenter: Binen Ye, Shanghai Normal University Tianhua College, China Abstract: Abstract—This study developed and verified a set of scales for evaluating the AIGC readiness of preschool education normal students. Through a systematic literature review and expert examination, the dimensions of the scale were determined. Exploratory factor analysis (EFA) was conducted using 112 pre-survey samples to establish the structure of the scale; Then, confirmatory factor analysis (CFA) was conducted on 385 formal survey samples using AMOS software, and a first-order structural equation model was constructed to evaluate the reliability and validity of the scale. The final scale consists of 5 first-level dimensions and 23 measurement items. The research results show that this scale has good reliability and validity.
4	13:45-14:00	IT3294	Developing Systematic Capabilities in New Engineering Education: Exploration and Practice Based on the “An-Yan” Intelligent Agent Author(s): Jie Chen, Guangchao Yu Presenter: Jie Chen, Anhui University, China Abstract: The widespread application of artificial intelligence in education has profoundly reshaped modern pedagogical practices, acting as a key enabler for the development of personalized and adaptive learning models. However, New Engineering Education poses unique challenges due to its complex knowledge structures, strong interdisciplinarity, and high-level problemsolving demands. Existing approaches often fail to capture the internal logic of engineering knowledge and lack sufficient reasoning ability, limiting their effectiveness in fostering core competencies. To address this gap, we developed the An-Yan model, an intelligent agent designed to cultivate students’ systematic capabilities by integrating structured knowledge with generative reasoning. This model precisely guides the learning process, facilitating systemic knowledge construction, interdisciplinary transfer, and enhanced problem-solving skills. Teaching practice at Anhui University showed that experimental classes using the An-Yan model achieved markedly higher scores in midterm and final exams compared with control groups, confirming its effectiveness. The study highlights An-Yan as a promising approach to meeting the complex demands of New Engineering Education and advancing the development of adaptive intelligent teaching systems.
5	14:00-14:15	IT3296	Gen-PBL in Vocational Big Data Education: Exploring Cognitive Offloading Behaviors via Structured GenAI Scaffolding Author(s): Shu Chen, Fenghua Liu Presenter: Fenghua Liu, Huzhou Vocational & Technical College Abstract: The integration of Generative AI (GenAI) into programming education accelerates code production but introduces pedagogical challenges regarding novice schema construction. This exploratory study investigates the behavioral dynamics of human-GenAI collaboration within a Project-Based Learning (PBL) environment, specifically examining how variations in instructional constraints influence learners' cognitive offloading behaviors. Through a 16-week quasi-experimental design involving 82 vocational college students, we longitudinally tracked two cohorts with divergent prior knowledge baselines as they interacted with GenAI under structured scaffolding (mandatory prompt templates) versus free exploration conditions. Our quantitative analysis suggests that the quality of cognitive engagement outweighs interaction volume. Hierarchical regression shows that prompt engineering quality is a significant predictor of project success ($\beta = 0.42$, $p < 0.001$), while raw GenAI querying frequency is not statistically significant. Furthermore, behavioral log data demonstrates that when structural scaffolds are abruptly removed, novice learners exhibit a distinct regression toward a "trial-and-error" regulation mode—marked by a high-frequency, low-quality querying loop and diminished project performance. These findings suggest that in GenAI-assisted environments, mandatory prompt structuring may serve as an external regulatory script. By intentionally introducing "cognitive friction," this approach compels learners to articulate problem decomposition before automated generation, thereby mitigating dysfunctional cognitive offloading. Based on these empirical observations, the study proposes design principles for preserving germane cognitive load through gradual scaffold fading and the integration of process-oriented assessments in future human-GenAI collaborative programming interfaces.
6	14:15-14:30	IT3235	Design and Evaluation of ICAP-Informed Prompt Scaffolding for Human–AI Collaborative Learning in Secondary Education Author(s): Xinyu Chen, Yihan Fu, Zhenxiong Zhang Presenter: Xinyu Chen, Tianjin University of Technology and Education, China Abstract: The educational value of generative AI (GAI) in classrooms increasingly depends on how students interact with these tools rather than treating them as “answer generators.” Grounded in the ICAP framework, this paper conceptualizes prompt scaffolding as an interaction-level constraint that translates engagement goals into implementable prompt requirements, shaping what learners must generate during human–AI dialogue without changing task goals. We instantiate this logic through two task-agnostic scaffold templates: an Active-oriented scaffold emphasizing procedural execution and verification, and a Constructive-oriented scaffold requiring explanation, alternative comparison, and boundary-condition reasoning. In a classroom feasibility study (N = 38; n = 19 per condition) on spreadsheet-function learning using a mobile GAI tool, we analyzed student-authored final task statements using an ICAP-aligned 0–2 ordinal rubric (Cohen’s κ = 0.797). The Constructive-oriented scaffold elicited substantially higher constructive engagement patterns than the Active-oriented scaffold (Cliff’s δ = 0.524). While raw posttest scores were comparable, covariate-adjusted analyses suggested higher adjusted posttest performance for the Constructive condition (η_p² = 0.27) given baseline differences. These findings provide initial mechanism-oriented classroom evidence that lightweight prompt constraints may help shape the expressed cognitive structure of human–AI interaction and offer a reproducible prompt-scaffolding approach for skills-oriented instruction.
7	14:30-14:45	IT3346	AI-Empowered Integration of Competition, Training, and Evaluation:Innovative Practices and Validity Analysis of the 2025 New Teacher Fundamentals Competition Author(s): Lu Sun, Lihui Wang, Lixue Zhou, Handan Dong Presenter: Lu Sun, Ningbo Education College (College of Education and Training), China Abstract: With the deep integration of the new wave of technological revolution and educational transformation, artificial intelligence (AI) is reshaping the pathways and paradigms of teacher professional development with transformative potential. As the new generation of educators, the pace and quality of professional growth among novice teachers directly impact the foundation of future education. Traditional teacher training models suffer from issues such as the separation of training and competition, highly subjective evaluations, and delayed feedback, significantly impeding the growth efficiency of new teachers. This study centers on the 2025 New Teacher Fundamentals Competition held in Ningbo, China. This competition pioneered the application of artificial intelligence as a core evaluation tool, constructing and implementing a novel teacher development model integrating "competition, training, and evaluation." This paper first outlines the theoretical underpinnings of the "competition-training-evaluation integration" model and its evolution in the AI era. It then delves into the specific construction and operational mechanisms of the AI-enabled model used in the competition, revealing how AI generates detailed diagnostic reports for each participating teacher through multimodal data analysis. Subsequently, through qualitative and consistency analysis of core data, it empirically examines the validity and value of AI-based evaluation. Findings reveal that AI-generated diagnostic reports demonstrate significant advantages in feedback immediacy, comprehensiveness, and personalization, effectively driving post-competition targeted professional development. However, data also indicates that current AI scoring shows relatively low consistency with human expert judges across certain dimensions, reflecting differences in evaluation criteria and emphasis. Finally, this paper summarizes the theoretical and practical contributions of this innovative model, identifies shortcomings in algorithm transparency and evaluation dimension balance, and outlines prospects for building a new data-driven, human-machine collaborative ecosystem for teacher professional development.
8	14:45-15:00	IT3252	An Empirical Study on the Influencing Factors of Student-Learning Outcomes Under Two GenAI Usage Behaviors for Machine Learning Course Author(s): Haibin Xie, Yijing Huang, Hanyin Zheng, Bixia Zeng Presenter: Haibin Xie, Yijing Huang, Guilin University of Aerospace Technology, China Abstract: Generative Artificial Intelligence (GenAI) is significantly impacting course learning of students in emerging engineering colleges. First, two specific GenAI usage behaviors are defined as GenAI-assisted deep learning and GenAI-substituted surface learning. Secondly, a measurement scale is developed to assess two GenAI usage behaviors and their link to learning attainment across four core areas: knowledge, methods, thinking, and professional competency based on Machine learning Course, and then confirms the scale's reliability and validity. Finally, a structural equation modeling (SEM) is constructed to explore the influencing factors of student-learning Outcomes for above two different behaviors. The result shows GenAI-assisted deep learning strongly supports learning attainment, while GenAI-substituted surface learning undermines it. Furthermore, these two GenAI usage behaviors are strongly and inversely related. The effect of GenAI depends not on whether it is used, but on how it is used. This provides clear evidence to guide students in using GenAI as a tool to build understanding, not as a shortcut to avoid learning.