This study investigates the effect of AI-generated feedback on university students’ essay writing proficiency in Hong Kong. By integrating generative AI into the revision process, the research examines how automated feedback influences the quality of students’ written work, as well as their engagement, motivation, and emotional responses during revision. The study employs a randomized controlled design, comparing students who received AI feedback on their essays with those who did not. Both quantitative and qualitative data were collected to assess the impact of AI feedback on writing improvement and student experiences. Quantitative analysis shows significant improvements in essay quality for students who utilized AI feedback, while qualitative findings highlight increased engagement and motivation, though students exhibited mixed emotional responses to the revision process. The results suggest that generative AI has considerable potential to enhance writing skills in higher education by providing timely, individualized feedback. This research contributes to ongoing discussions on the role of AI in education, particularly in supporting language instruction and student development.

This study developed and validated a Computational Thinking-Project Based Learning (CT-PBL) framework to foster sustained interest in software education among software (SW) gifted students. Despite rapid expansion of SW gifted programs in South Korea, maintaining student interest remains a critical challenge. The framework integrates Dewey's experiential learning theory with Hidi and Renninger's four-phase interest development model to systematically support students' progression from situational to individual interest while developing computational thinking competencies. The framework was developed through systematic literature analysis, expert review, and field practitioner validation, demonstrating strong content validity. External validation employed a quasi-experimental design with 16 SW gifted classes over 12 weeks. The experimental group showed significant improvements across all interest development stages compared to the control group. The CT-PBL framework contributes to both theory and practice by empirically validating interest development in computational thinking contexts and providing immediately implementable tools including a structured 12-week protocol, teacher guide, and student portfolio system. These findings have important implications for educational programs aiming to foster both technical competence and sustained engagement. While acknowledging limitations related to the Korean educational context and quasi-experimental design, this study provides a robust foundation for nurturing future digital innovators through interest-driven computational thinking education.