Following its launch in 2021, this workshop seeks to gather fresh insights into employing Artificial Intelligence (AI) systems and intelligent tutors for education and learning through the use of multimodal data sources. The Multimodal Artificial Intelligence in Education (MAIEd) workshop builds upon the CrossMMLA workshop series at the Learning Analytics & Knowledge Conference and the Multimodal Immersive Learning Systems (MILeS) at the European Conference of Technology Enhanced Learning. The MAIEd workshop calls for new empirical studies, even if they are in their early stages of development. It welcomes novel experimental designs, theoretical contributions, and practical demonstrations that can prove the use of multimodal and multi-sensor devices “beyond mouse and keyboard” in learning contexts for automatic feedback generation, adaptation, and personalisation in learning. Through a call for proposals, we seek to engage the scientific community in opening up the scope of AI in Education towards novel and diverse data sources.