Faculty at research and professional schools are adopting new assessment methods as generative AI blurs the line between student work and authored content. At NYU Stern, data‑science professor Panos Ipeirotis moved to AI‑assisted oral examinations after written assignments began to read like consulting memos but lacked depth of understanding. Concurrently, higher‑education analysts and practitioners are shifting expectations: 2026 may mark a pivot from novelty AI use toward tools that demonstrably deliver ROI for teaching, advising and administrative workflows. Institutions face trade‑offs between academic integrity, assessment design and investment in validated AI tools.