AI arrived on campus as a crisis but has since become banal. Students can now do basically all their schoolwork automatically with AI—answering scientific problem sets, writing computer code, and drafting essays. Professors like me have no idea whether we are really teaching or just LARPing for students who are also pretending to occupy their role. Worse, some of us are offloading the drudgery of writing recommendation letters, grant proposals, and even scientific papers to AI. But at the same time that AI fouled up the student-teacher relationship, the technology benefited other corners of campus. Researchers quickly discovered that AI is enormously useful for knowledge production, especially in the natural and applied sciences: in protein-structure prediction, materials design, and drug discovery, for example. Pressured, in part, by donors, boards, employers, and parents, college administrators have announced AI initiatives, invested in AI infrastructure, and established AI degree programs. These two forces—students using AI to their detriment and colleges using it to their benefit—might seem to be in conflict. Students are supposed to do everything longhand to …