In about three years, generative AI has become a normal part of design work. Architecture students in India already use it often, but most schools still have no clear rules for how it should be used. This paper brings together research on three questions: how AI is changing architecture education, why students submit AI-generated work, and which institutional responses actually help. Studies show that AI is most useful at the start of the design process, where it speeds up idea generation and helps produce many polished visuals. It is much less helpful in later stages that need critical judgement, technical accuracy and real originality. Whether AI improves or harms learning depends mainly on how teachers frame and control its use. The evidence on copying is clear but uncomfortable. Current AI-detection tools are unreliable and often label the writing of non-native English speakers as AI-generated. In India, this makes them both unusable and unfair. Experienced teachers also cannot reliably tell AI-generated work from student work just by reading, and giving “real-world” project briefs does not, on its own, stop students from using AI. The research points instead to a set of simple, practical measures: clearly stating how much AI use is allowed for each task, grading the design process and not only the final output, designing assignments that AI cannot complete well, requiring students to declare which tools they used, writing a code of conduct together with students, and reducing workloads and deadline pressure that encourage plagiarism. The paper ends with a year-by-year plan for integrating AI across the five-year B.Arch programme and highlights key topics for future research in India.
Keywords: architecture education; generative AI; academic honesty; assessment; design studio; AI literacy; India; Council of Architecture
