This paper examines the current state and challenges of generative AI literacy in Japanese higher education through classroom observations at Doshisha University and international comparative analysis. Students extensively useChatGPT for translation, assignment completion, test preparation, and idea generation, with 70-80% openly acknowledging AI translation use.However, usage patterns vary significantly, from constructive iterative translation supporting learning to mechanical copying that bypasses cognitive engagement.Faculty face challenges including academic integrity concerns, assessment redesign needs, insufficient preparation, and culturalinstitutional factors such as consensus-oriented decision-making that delays individual policy implementation.International comparison reveals that universities in the United States, Europe, Australia, and Singapore have evolved from blanket prohibition toward nuanced policies distinguishing appropriate from inappropriate use, though approaches diff er in implementation speed, authority distribution, and balance between enforcement and education.Japanese universities can develop culturally responsive AI literacy approaches by drawing on
Anthony LaVigne (Wed,) studied this question.