Could we teach a computer to lie?

[Angrboda]

Compatibilist and Determinist conceptions of free will have no obstacles to being implemented on a computer as they, ultimately, simply model 'will' as the outcome of a calculation based on relevant factors input into the decision. Oddly enough, libertarian free will does something similar in modeling will as the confluence between personality (the soul) and the relevant reasons and such that apply to the choice under consideration. The only sense in which this differs from compatibilist and determinist conceptions of free will is in asserting that the behavior or traits of personality aren't themselves deterministic, but rather are the result of an indeterminately free will. If that latter is the case, then a computer likely could not model that form of decision making. It's somewhat implicit in the very definition of such libertarian free will that it cannot be modeled by a deterministic process. Which, in some ways, insulates it from any attempted disproof. At bottom, then, libertarian free will becomes an assertion not appealing to any facts. It's true that by definition it could not be modeled by a computer, but things that are only true solely through definition and assertion are not particularly interesting, nor particularly likely to be true.

I explored what I think is a relevant model for our will, here, which provides a process which can readily be implemented by a computer.