[JesseK]

To me, there are two fundamental problems that no sub-set of current 'laws' can address.

1.) In the current state of the art, robots are to individual humans as corporations are to individual businessmen/women; that is, in addition to the benefits they are also a way to make behavior which we may otherwise feel guilty about pseudo-anonymous. Drone strikes, high frequency trading & information theft are current examples, whereas general inter-human conflicts on a smaller scale could be at stake in a larger A.I.-driven society. It should be noted that 'robots' in this sense do not constitute only physical machines -- something that I don't know whether Asimov understood at the time.

2.) We (humans) have yet to fathom a viable 'end-game' for a fully autonomous society that we're comfortable with because we're afraid that with autonomy will come an eventual sense of willpower. In the epic struggles of many centuries ago, how often was the willpower of the many weak able to overcome the brute strength of the powerful few? (hint: the answer lies in the beginning of most U.S. history textbooks) Thus, most scenarios we perceive involve humans pulling along a robot-ridden carriage in a distant future -- and we all know what happened to horses when the automobile was invented.

Not only would any 'law' have to address humans creatively destroying each other (with increasing creativity might I add) but the set of laws would have to address our desire to maintain dominance as a species.