More colleges are turning to oral exams and analog assessment practices to reduce AI-enabled assignment misconduct. Reporting described oral defenses in Cornell University courses and typewriter-based writing exercises that remove laptops, chatbot access, spellcheck tools, and other digital supports. The approach shifts assessment from AI-completion to demonstrated reasoning and live communication, aiming to make it harder to outsource work while preserving learning objectives. Faculty reported using these methods both as a deterrent to AI-generated submissions and as instruction in writing processes. As AI tool access becomes ubiquitous, these assessment changes show institutions testing low-tech integrity safeguards alongside policy updates and AI transparency instruction.