Only a minority of higher‑education institutions have formal AI governance; education leaders are racing to build policies defining acceptable use, data privacy, and academic integrity as generative models proliferate. Guidance documents emphasize procedures for development, deployment, and oversight to manage bias and maintain learning outcomes. At Columbia, a vice dean for AI initiatives publicly cautioned against unguarded adoption—arguing that practical harms and academic misuse have outpaced institutional guardrails. Administrators are piloting narrow, use‑case pilots and recommending small pilots, strict guardrails, and staff training before scaling campus‑wide. Campus leaders should prioritize AI governance frameworks tied to pedagogy, student privacy, and research integrity—embedding ethics, auditing, and clear escalation channels into procurement and syllabus practices.