I don't see this being off.
If you understand the enemy behavior better by use of models, then the model will have been valuable.
But , I don't think that any enemy we might ever have will be totally free of intelligence.
As we learn them (including with models) they will also learn us.
Thus the value of older learning or seldom updated models will degrade as a function of how well we are understood.
"everyone has to eat, sleep, defecate, drink water, and so forth. You turn that against them."
Perfectly true, especially the "and so forth" because there are a large number of behaviors that you can predict a human being will do.
But this is no less true in reverse , because we are just as human.
In the worst case scenario some of the enemy could be genuses in possession of computer power and even worse, a copy of our model.
There is a general principal here.
As in the drone that Iran recently captured , any tecnology or tecnique that we rely on too much has a potential for becoming a trap.