They're going to have change everything so use of an AI assistant doesn't matter because once they graduate they're just going to continue using it anyway.
If it's a math for finance course then some kind of model building for the midterm and being marked on the quality of the model or something. If AI becomes so good that it always chooses the best fitting model and requires no numerical optimization then they will have to change the courses to be more like UChicago where it's primarily undergrad directed research but AI assisted.
Why teach kids how to read, they can just take a picture of whatever text they want and have AI say it aloud.
This is a dumb take. It's like not teaching kids 1 + 1 because a calculator can do it for them.
The challenge I think is that students then struggle because they used AI throughout the semester and didn't actually learn. The proper response would be to be strict and fail students that don't perform to a satisfactory level, but this messes with the funding incentives.
You can only lead a horse to water, you can't make it drink. Maybe a student's sincerity should play a larger role in the admission process, maybe with a sharp expense curve such that students judged to be more sincere have to pay less tuition. It is an inherently subjective evaluation though.
Edit: I completely misread your comment. Asking students to build a model is not a finance class anymore.