世界模型驱动未来教育变革:基于"四未来"要素的机制阐释与形态塑造OA

World models are advancing artificial intelligence from symbolic intelligence toward embodied intelligence,address-ing some of the limitations of large language models in perceiving the physical world,modeling causal relations,and supporting em-bodied practice.Drawing on educational principles and the technological affordances of world models,this study first develops an en-abling mechanism centered on the cultivation of core competencies,explaining how world models may support learners' cognitive,ac-tion-oriented,social,and affective development.It then elaborates the internal mechanisms through which world models may con-tribute to the reconstruction of educational ecosystems,including dynamically evolving embodied learning environments,teacher-stu-dent-machine-environment collaborative interaction,spatialized instructional resources based on causal relations,embodied pedagog-ical models,and developmental assessment supported by prospective simulation.Furthermore,this study identifies practical pathways through which world models may reshape four key elements of future education:future teachers,future classrooms,future schools,and future learning centers.Finally,it examines potential risks associated with the educational use of world models,including issues relat-ed to educational equity,cognitive development,and privacy protection,and proposes recommendations for their sustainable applica-tion in future education from the perspectives of adaptability,collaboration,and ethics.Overall,the educational application of world models creates new possibilities for the deep integration of artificial intelligence into educational processes and offers new insights into the systemic transformation of key elements of future education.