Yerin Seung, James D. Basham, Taehyun Kim, Jennifer L. Lohoefener
ABSTRACT Background Artificial intelligence (AI) can support more personalised learning in K‐12 education, reshaping the educational experience for all students. However, little is known about how AI is being designed and implemented to support accessible and equitable learning experiences for all students, including those with disabilities and academic challenges. Objectives This scoping review investigated how AI has been integrated in K‐12 educational settings and what types of AI, devices and modalities are used to support student‐AI interaction of diverse learners. Methods We identified 70 empirical studies that included student‐AI interaction in K‐12 education between 2014 and 2023. The review focused on the scope of literature, types of AI and devices used, and the modalities used for student‐AI interaction in K‐12 education. Results and Conclusions The number of empirical studies on AI integration has surged worldwide in an interdisciplinary fashion across school levels and subjects. However, only a limited number of studies reported the diversity of students. Chatbots and tutoring systems were the most often used AI tools, computers and tablets being the most often used devices. Research across diverse academic backgrounds commonly adopted a single modality to interact with AI. At the same time, multimodalities were more prevalent for the AI's output, both of which were text‐dominated. We discussed the necessity for interdisciplinary collaboration and adopting multimodal AI to enhance accessibility for students with diverse backgrounds.