Free Will might be nothing more than a trick the brain plays on itself

Free Will Is an Illusion, but You're Still Responsible for Your Actions -- http://chronicle.com/article/Michael-S-Gazzaniga/131167/

Unless one believes in dualism, I can't see how one can believe that there is such think as free will.

Suppose, our mind is indeed designed to have a postdictive illusion of choice, does it make sense to equip an AI with it?

Does an AI need to feel in control? What kind of role can an illusion of choice have in reinforcement learning? Wouldn't it introduce errors on smaller time scales, e.g. when catching a ball? Or in building up procedural memory in general?

so if my brain made me think i chose to pick up a box of cereal, after the fact, what made me pick it up?

Philosophy has been tackling this issue forever. And, in my opinion, plenty of brilliant minds have provided excellent cases throughout philosophical history that make it clear we most certainly do not have free will. Oh, we certainly love to feel and think that we do. Primarily, this provides us a sense of control over our lives, over the space we inhabit across a given time period. We also love to punish others as if they did, as free will is the foundation of determining and adjudicating moral responsibility for one's actions. But that's more a reflection of how we want to see the world operating according or in opposition to our whims or, as the article puts it, a product of tricks our brains play on us that helps us feel like we have control.

The problem is we have no satisfactorily logical and coherent case for explaining how a human agent exerts controlling causality on itself and the space around it through time. Because of this, most attempts at defending/arguing the existence of free will typically hinges on altering the definition of freedom that must be obtained in willing, an effort to carve out a defensible position that looks like free will, but is a rather amusing game of endlessly moving the goal posts toward a compatible definition of freedom.

For compatibilists, this pretty much always turns into a logical playground of arguing that being able to do otherwise would be the case if some increasingly creative number of things about oneself had been different. Peter van Inwagen's replays-by-god are a pretty amusing example here. Compatibilists accept causal determinism outside the human agent, but seek to carve out a space within for the human agent to will and act freely.[1]

Incompatibilists, naturally, reject the notion that freedom and causal determination can co-exist. So they typically argue over a number of varying-strength propositions that seem to more often accept agent freedom as a given, but causality as the thing worth arguing over.

For some time yet, I imagine we will continue to argue over and shift the lines of where the human will may be considered the causal source of human action. But I don't think we're ever likely to come up with a definitive argument that is intelligible enough to be satisfactory.[2]

---

[1]: There are, of course, some who like to differentiate between acting and willing freely, of course.

[2]: I can't help but wonder where this conversation would be if we hadn't had many intervening centuries of religious dominance in various human endeavors, including philosophy, that sent us on centuries of building gods into our arguments, and then subsequently taking gods back out and winding up nowhere closer to a decision. Well, maybe a bit closer, but we're still dealing with the implications of gods on human will and action in the 21st century.

nah, quite a relief after reading this, there must be some other subject to blame, for all my wrongdoings.