Karl Popper once said:
I do believe in freedom and reason, but I do not think that one can construct a simple, practical and fruitful theory in these terms. They are too abstract, and too prone to be misused; and, of course, nothing whatever can be gained by their definition.Karl Popper in “From the archives: the open society and its enemies revisited” found here.
In fact, this is one of the reasons why libertarians politics — which is formulated as a theory about ‘freedom’ — is incompatible with Popperian epistemology. 
The problem with words like “freedom” and “reason” are, first, that they are quite vague and, second, that they are ‘religiously loaded’ terms. They are vague because they are words that point to many different ideas and try to place them under a single umbrella term. The end result is that they co-mingle multiple concepts as if they were a single unified concept. There are many conflicting ideas about what should count as ‘freedom’ just like there are conflicting ideas about what should count as ‘fairness.’ We shouldn’t expect such terms to ever have a single essential meaning.
But the bigger problem is that these terms are “religiously loaded” terms that people feel an overwhelming need to ‘own.’ People are constantly trying to capture the terms for political reasons because of the halo of moral connotation attached to such words.
If your goal is to come up with a hard-to-vary explanation that makes your ideas easily testable (and thus improvable) then these terms are just not going to be helpful. Put simply, if what you care about is actually maximizing ‘freedom’ then your first goal will be to dispense with that word and find a better way to frame problems that will allow you to formulate your theory in a more testable way.
I’ve criticized the use of the word “coercion” in the past for exactly the same reason. You can’t form a good political theory of terms like ‘freedom’ and ‘coercion.’ In fact, part of the appeal of using such terms is exactly that they are subjective terms that make your theory impossible to test or refute — a feature valuable for political internet debates but counter-valuable for explanation-making.
The simple truth is that I feel the same way about the term ‘Free Will.’ I do believe in ‘free will’ — at least in the one kind of free will allowed for by our current best-known theories. But I don’t see how such a vague and religiously loaded term could ever turn out to be useful in formulating theories about intelligence, AGI, or really anything.
Yet I find minds much greater than mine are constantly desiring to figure out how to reconcile free will with current theories despite not being a productive way to go about framing the discussion in the first place. Or so it seems to me. But given the importance the term clearly holds to so many individuals, I suspect we’re nowhere close to recognizing the value of just… well, dropping the term altogether when trying to formulate good explanations. I doubt a theory of AGI will in any way reference “free will.”
If you must speak of ‘free will’ I suspect the better approach is to work out what the implications of it are given our current best theories. What does it mean for free will to be deterministic? How does that, if at all, change our perception of it? To me, that’s the more productive approach.
 Oh, let’s be honest. Liberals and conservatives aren’t really doing any better on this front, constantly trying to define their politics in terms of words like ‘freedom’ and ‘reason.’ So no harm no foul specific to libertarians here.