This study investigates how universities are developing normative frameworks to regulate the use of generative AI tools in higher education, with a particular focus on balancing empowerment and discipline. Drawing on theoretical lenses such as Foucault’s discipline theory and contemporary AI ethics, the paper analyzes policy documents from institutions including Harvard, Oxford, and several top Chinese universities. Using Latent Dirichlet Allocation (LDA) topic modeling, the study reveals five dominant governance themes—ranging from academic integrity enforcement to pedagogical empowerment. The findings highlight a global shift from restrictive to balanced, ethics-informed AI governance, with significant disciplinary variations. The paper concludes by proposing a “principled permissiveness” model that combines transparency, accountability, and pedagogical innovation in future AI governance.



