A liberal‑arts college’s experiment to embed and evaluate large language models in teaching has spotlighted both institutional ambition and student concern. Inside Higher Ed’s coverage describes campus pilots that compare and assess multiple LLMs for learning and research use; the move is part of a broader trend of smaller institutions adopting curated AI tools rather than ceding digital learning to third‑party platforms. Students and surveys, however, underscore rising anxieties: a Project Tomorrow report and other polling show that students’ top AI worries include misinformation (69%), weaponization of AI to harm others (65%), data privacy (57%), inaccuracies in outputs (56%), and fears of being falsely accused of cheating. College leaders now face a dual task: deploying AI to enhance pedagogy while building safeguards, honor‑code protocols and data protections to address those concerns.