ChatGPT on campus: how top U.S. universities govern generative AI across higher education

Generative Artificial Intelligence (AI) tools have transformed the landscape of higher education. U.S. universities often develop policies in a reactive manner without a knowledge-base or with minimal understanding of the approaches undertaken by other institutions, leading to unintended governance gaps. Our analysis spanning the top 50 U.S. universities finds that most universities adopt a restrictive or centralized approaches across key dimensions, including the default prohibition of AI-generated content in academic integrity policies, explicit guidelines for AI use in instructional communication, prohibitions on inputting sensitive data into AI tools, mandatory disclosure of AI use, and structured revision processes for AI policies. While most institutions favor a top-down approach, some implement innovative and community-based policies emphasizing continuous engagement, inclusivity, and critical thinking. Building on these insights, we propose four recommendations: shifting towards flexible approaches, exploring community-based strategies for AI policy development, balancing flexibility with caution to meet educational and research needs, and embdarcing AI-integrated classrooms. Additionally, we identify four gaps: aligning AI policies with evidence-based pedagogical approaches, establishing a unified dissemination strategy, developing policies that regulates and promotes AI tools for personalized learning, and practical implementation challenges.