A new Instructure poll found widespread AI use in higher education and classrooms: 90% of college students and 61% of instructors reported using AI at least occasionally, but only 11% of instructors said they received comprehensive AI training. The survey released Tuesday highlights a gap between tool adoption and educator preparation, with many institutions still setting boundaries and support structures. In parallel, new research out of the Centre for Economic Policy Research reported a learning tradeoff in China: AI boosted homework scores but later exam performance fell. The study analyzed 26,811 students and found exam scores dropped after AI adoption, with persistent impacts over time—especially among students who “outsourced” homework completion quickly. The developments create a practical compliance and pedagogy challenge for colleges: AI use is already normalized, but the sector is still determining how to align course expectations, assessment design, and academic integrity policies without relying on assumption that AI use automatically benefits learning. For student success leaders, the near-term focus is likely to shift toward educator training, AI use guidance, and assessments that test learning processes rather than shortcut-driven output.