Researchers explore classroom applications of AI new learning platforms Sadie Harley Scientific Editor Andrew Zinin Chief Editor Artificial intelligence (AI) is increasingly embedded in classroom practice in Hong Kong, following the Education Bureau's release of the Blueprint for Digital Education Development in Primary and Secondary Schools and its commitment to supporting innovative technology in schools. The latest research from The Education University of Hong Kong (EdUHK) underscores that AI's educational value lies not in computational speed but in its integration with rigorous, evidence-based and student-centered frameworks for assessment. These findings provide timely guidance for Hong Kong's digital transformation in education.

Led by Professor Zi Yan of the Department of Curriculum and Instruction, the research team analyzed data from 151,969 students across 19 countries and regions in the Program for International Student Assessment (PISA). The study examined the impact of core dimensions of formative assessment on learning outcomes. Results indicate that frequent feedback alone does not necessarily enhance learning.

Instead, students' reading achievement improves when teachers clarify learning goals, monitor progress systematically and adapt instruction in response to assessment evidence. The work is published in the journal Education Sciences. Yan said, "An assessment ecology is shaped jointly by school policies, teacher practices and teacher-student interactions.

In an examination-oriented system, teachers may devote considerable time to feedback, yet students may not use it effectively. AI's potential depends on designs grounded in sound theory and empirical evidence. "Programs can improve learning only when they incorporate clear learning goals, ongoing assessment and support for instructional adjustment." He emphasized that assessment for learning must be firmly grounded in local realities, including large class sizes, teacher workload and the need for school-based support.

It cannot be achieved simply by importing policies and practices from elsewhere. Building on these insights, the team has developed two platforms since 2024: - EASE (Efficient Adaptive System for Education)—Integrates formative assessment, self-directed learning strategies and AI. Adaptive algorithms generate personalized learning pathways, while real-time feedback, automated marking, error analysis and learning reports support informed decision-making by teachers and students.

- FAITH (Formative Assessment Innovative Teaching Hub)—Provides authentic classroom examples and subject-specific materials, combining teacher-led assessment with active student participation. FAITH helps students move from passive reception to active use of feedback while offering professional development resources and enabling schools to build a shared language and body of practice around assessment. FAITH currently has around 350 registered users, while EASE has more than 2,200.

Teachers report that the platforms enable personalized learning and more responsive instruction; students highlight increased motivation and engagement in self-directed learning.