This exploratory study presents “AI for Universal Design for Learning (UDL)” as a core principle in defining the role of artificial intelligence in education. It argues that generative AI should be regarded as a stepping stone and an opportunity to enact a completely different vision of education rather than merely a means to achieve existing standardized educational goals. To this end, the study (1) critiques the AI-related discussions centered on productivity, efficiency, and economic utility, which reinforce the normativity regime shaped by sociotechnical acceleration; (2) examines how AI can support and enhance the guidelines of UDL in literacy education; and (3) delineates the role and limitations of “AI for UDL” and discusses the accompanying principles of “mutual embodiment for learning (MEL)” and “critical meta-literacy” as necessary for effecting transformative change. Through this line of work, the study aims to establish a conceptual framework for pursuing changes in educational practices in light of the rise of generative AI and seek to broaden the horizons of critical, creative, and relational literacy education.