Actually, the big thing about the chimp experiment(as well as the tongue input ones) is it proves this conception is wrong.
As I said, the chimp brain not only controlled the arm(as opposed to flailing it about randomly) , it developed sections of the brain dedicated to control of the robot arm.
On the human level we've set up non-invasive rigs, but they aren't very sensitive and just moving a mouse cursor-type object with them is a chore, much less operating in a full 3D space with varying levels of "click".
And we're actually born with very little wiring.
That's why babies do a lot of nothing for a while. They don't know how to work their bodies.
But we've found out recently that we can learn new things well into adulthood(albeit slower than a child does).
A person born deaf can learn to hear if he's given a way to pick up sound. Someone born blind can learn to see if given a way to recieve light.
Or you could just wire the arms into an adult and let the incredible adaptiveness of the human brain figure it out, as it will based on what we know now.