Well most of the explanations I personally am interested in have a public and therefore high epistemic value. Since they are published and repeatable. Certain behavioral tests or self-report surveys on the other hand have lower value because of the private nature of what were intending to test. They still have some value though.

Private experiences may be very important for individuals and like I said elsewhere, free-will is certainly a useful idea in religious contexts. I do not believe this has any bearing on practical matters such as our ability to predict or understand the actions taken by any given system such as humans or computers.