Morosoph's Journal: An Implication of the Zeroth Law of Robotics

Most of us will be familiar with Isaac Asimov's Laws of Robotics, which, by simple virtue of their popularity, are bound to have an influence upon robot design.

In practice, however, human nature being what it is, we are bound to create war machines. Even if we do not, we should make ourselves aware of an implication of the zeroth law:

A robot may not injure humanity, or, through inaction, allow humanity to come to harm.

The implication is this: any robot that is created that does not follow the laws of robotics, should not be able to reproduce. If it needs to build robots, in turn, they must have a lower complexity than itself, and carry the same condition. Further, a robot that can create robots and does not carry this rule should be destroyed.

It's pretty clear that in the event of war, the three original laws would have to be suspended for the sake of the zeroth, further, having made a strategic decision with (one would assume) a higher-powered machine, that strategic decision should not be overruled because the robots on the ground lack information. Although a better solution might be that war robots carry a "degree of confidence" in their creator robot, or the strategist, this ignores the human hand, where humans will not want robots to be surrendering on their behalf. Humans believe in their wars, and won't want robots contradicting them.

So, why is this implication so important? The answer lies with genetics. The reason that we put so much effort into survival and the protection of our families, and to a lesser extent, our friends, is that we are genetically programmed that way. Those genes that have enhanced our survival qualities have been successful. Unlike most of us, robots can truly be selfless.

However, if robots build robots, mutations are bound to occur. If it were left unchecked, one would find that those mutations favouring the robot's own survival (and those of its 'offspring') would prevail over robots that do not have that mutation. Robots would slowly become selfish, and promote their selves and their own 'relations' above human beings, for at some point, the wired-in laws, insofar as they existed in the first place, would be subject to mutation. Thus reproduction-capable robots not carrying the laws must face capital 'punishment', as must those disobeying them (since the laws could remain intact, yet somehow be "wired around"). Asimov's laws will then be part of the survival criteria, the only exceptions being those robots that cannot reproduce, and therefore never get the chance to evolve.

With Asimov's laws as part of the survival criteria for reproducing robots, we will find the laws being respected and strengthened down the robot generations, for the criteria for robot survival will match the criteria for robot self-propagation.