How AI-assisted MBA projects can be evaluated fairly

Educational institutions must establish secure, campus-wide artificial intelligence environments where faculty can evaluate how state-of-the-art AI models respond to assignment prompts before students receive them. This approach, similar to systems at Harvard and Wharton, allows educators to better understand potential AI-assisted outputs and design more robust evaluation criteria.

Such controlled environments enable institutions to anticipate how students might use AI tools and develop assessment methods that remain meaningful in an AI-augmented learning landscape. By testing assignments against advanced models first, educators can refine rubrics and ensure that evaluations measure genuine learning outcomes rather than penalising or inadvertently rewarding AI use.