As a small country with great ethnic diversity, Singapore embraces inclusive education and is committed to optimise the potential of every student. One recent advancement is the implementation of full subject-based banding that allows lower secondary students to learn in one of the three levels that match their strengths and interests in respective subjects. It makes learning more relevant to students’ learning needs in each subject and provides meaningful interactions among diverse students who were previously streamed based on their ability levels, creating opportunities for students to appreciate diversity in an inclusive environment. Yet, teachers often feel unprepared to optimise learning of diverse students (Florian and Pantić, 2017), constrained by their knowledges, skills and workloads (Huang, 2022). Artificial Intelligence for Education (AIED), when developed and used in equitable and inclusive ways, has the potential to help bridge social and learning gaps when moving towards inclusive education. For example, Adaptive Learning System (ALSs) allows students to learn procedural knowledge, at an individualised pace with personalised support. ALSs also have the potential to identify a particular student’s learning difficulties so that interventions can be implemented within and outside ALSs. AI-powered assistive technologies such as speech recognition can also be included to support students with impairments. However, there are limitations to using AIED for inclusive education, if the technology itself is not designed to be inclusive and equitable. Inclusive education often requires changes in teaching methods, school culture, and policies to ensure that all students have equal opportunities to succeed. These changes are hard to be addressed solely by using AIED alone. In this presentation, we adopt Cerna, et. al. (2021)’s conceptual framework to propose five aspects of consideration in adopting AIED for inclusive education. First, for AIED to meet diverse learning needs, the designing of AIED needs to be in collaboration with various educational stakeholders so that AIED is human-centred and reflects student diversity. Second, resources need to be provided for equal access, particularly by students from low-income families. Third, teacher capacity should be developed, particularly for teacher-AI augmentation in inclusive classrooms so as to avoid AIED depriving student’s agency and voices in learning. Fourth, processes and outcomes of inclusive education and students’ achievements should be rigorously evaluated for continuous improvement. Last, culture and governance for both AIED and inclusive education are needed to sustain the effective use of AIED for inclusive education.

Cerna, L., Mezzanotte, C., Rutigliano, A., Brussino, O., Santiago, P., Borgonovi, F., and Guthrie, C. (2021). Promoting inclusive education for diverse societies: A conceptual framework. OECD Education Working Paper No. 260. Accessed 27 January 2023 https://www.oecd-ilibrary.org/docserver/94ab68c6-en.pdf?expires=1675088838&id=id&accname=guest&checksum=31B1AA71231B12EAE94C3C1AE4963CBC Florian, L. and Pantić, N. (2017). Teacher education for the changing demographics of schooling: policy, practice and research. In Florian, L. and Pantić, N. (eds) Teacher Education for the Changing Demographics of Schooling: Issues for Research and Practice. Switzerland: Springer. pp. 1-5. Huang, J. S. (2022). Optimizing the potential of every student in Singapore - Is AIED a solution to embrace the complexity in teaching and learning? In E. Walton and R. Osman (eds) Pedagogical Responsiveness in Complex Contexts. Switzerland: Springer. pp. 75-94.