[Synapse elist] warbots

[Leonel Moura]

Most probably robots will evolve in another direction distinct from 
human intelligence. Many animals did that and it works fine. If the 
(human) intention is to achieve the most possible autonomy for 
robots, then mimicking human intelligence, behavior and skills is not 
the best approach. First, we don't know how we work. Second, 
considering the human model is to aim at the run out of a process and 
not to trigger one. In fact this is a formalistic proceeding because 
it tries to implement very specific and predetermined features, and 
not to let the behavior emerge. Third, to do that one needs a lot of 
control. And as everyone should know by now control by itself doesn't function.

Military robots can perform many tasks, from surveillance to random 
carnage (like cluster bombs do), but once they must fulfill some kind 
of protocol based on human-like decisions many errors will be made. 
Keep in mind that it happens frequently that militaries are killed by 
the so-called friendly fire.

Anyway war and war crimes is one and the same thing. It is an 
obsolete way of solving conflicts.

That is also why I am much more interested in considering robots as a 
new species in the planet, with their own life, behaviors and skills. 
What does not mean that conflictual situations between humans and 
robots will not rise. But people are often killed by crocodiles, bees 
and many thousands by bacteria as well.

Robots will just add some more stress to the human environment. The 
kind of stress that enhances evolution.