I'm going to assume you're giving this hypothetical student the benefit of the doubt, so we can discard the "ACCACA DABACCA" version of test prep (Which is what the professor in the linked article is alleging)
https://youtu.be/0cyS4HZ_Dxc?t=56With that out of the way: Why is it bad for a student to study for a test from an unofficial test bank?
It's really stinking hard to write a good test question. It needs to be possible, clear, and not exactly* like problems you've done in class and homework. There are a few exceptions to this for "canonical" problems/derivations. Capacitance of parallel plates, for example.
It's very different if a student chooses a problem to study, and that problem is on the test, versus a student who is given a problem to study that is from the test. For example, if I give a "derivation of cylindrical capacitor" in my test prep packet, and a student looks at that and also studies the spherical and parallel plate geometries, that's great for them! They thought deeper about what was going to be on the test, and directed their attention accordingly.