Ralph Mann (USATF sprint biomechanist) says that elite level sprinting is not natural in the mechanical sense that it's not something you could come out of the womb doing. While I don't agree with many of his conclusions I do agree with this one. He mainly cites the very active recovery of the push-off leg and the high knee positions observed in elite sprinters as things that in no way happen naturally. When you look at the best sprinters in the world they toe off much closer to bottom-dead-center and the thigh is more quickly actively recovered following toe-off. While the first point can at least partially be explained by their greater power / impulse (they get off the ground quicker) the latter point is clearly due to active hip flexion. Also, I've never seen a child run with very high knees and a posteriorly rotated pelvis as in elite (of the elite) sprinters.
From a motor learning standpoint things get more tricky. We want athletes to run without thinking. Their ideal technique should be developed in practice such that it comes automatically…or naturally. This is when performance will be maximized. On another motor learning point, we often have to coach / reeducate things that probably would have come natural (out of the womb) but have been reverse-coached over time. For example, 'kick your butt' is a common sprinting cue for coaches who don't know any better. If this becomes natural (automatic) over time then we might have to coach them back to where they were prior to the faulty coaching cues. The same holds true for athletes who have their own misconceptions on technique that have manifested themselves in poor technique over time.