Colleges are redesigning assessment to accommodate widespread AI use. The University of Sydney implemented an assignment menu that permits AI tools for much work outside class while preserving in‑person assessment, aiming to preserve learning outcomes and reduce cheating. Cornell announced a discipline‑independent module to build critical thinking tailored to the AI era, providing faculty a framework to integrate those skills across courses. These experiments reflect an industry shift from blanket bans toward calibrated academic design: alternative assessments, in‑class demonstrations, and scaffolded assignments that emphasize process over polished output. Institutions adopting such models hope to protect academic standards while preparing graduates to use AI responsibly. Administrators should expect pressure to scale these pilots, update faculty development, and revise honor codes and assessment rubrics.