Faculty across campuses report that AI‑detection tools are changing student behavior: some students now use generative AI specifically to avoid stylistic features that detectors flag, while nonnative English writers are being disproportionately misclassified. Researchers and instructors warn that detection tools are reshaping writing into a performance to be managed rather than a practice to develop. The trend has led to lawsuits and campus policy disputes over fairness and academic integrity. Colleges face a policy tradeoff: deploy blunt detection tools and risk bias and defensive behavior, or adopt pedagogy and assessment redesign that explicitly integrates AI into learning objectives and evaluation.