The shift toward oral evaluation is spreading beyond single-course experiments as instructors report that take-home writing can look AI-perfect while students struggle to defend what they submitted. The trend is driven by faculty workshops, classroom policies that restrict AI use on certain work products, and pairing strategies that combine written output with real-time student explanation. At the University of Pennsylvania, for example, the Center for Teaching and Learning is supporting faculty training so instructors can use oral formats to verify understanding. Cornell’s approach similarly removes technology from the assessment itself, forcing direct interaction between student and instructor. For campus policy leaders, these developments raise immediate implementation questions: grading consistency, accessibility accommodations, and how institutions document assessment integrity without treating student performance as automatically suspicious.