People are already robotic enough. We don't need actual robots putting us to shame.
A counterpoint to those that believe robots "just wouldn't do war": Pride isn't a non-negotiable factor in beginning war or large scale hostile action. To well developed AI, it may be pertinent that humans be used much in the way lab rats are. Or forced labour. Robots can't be taught ethics, they can be programmed to become more sensitive to context through repeat exposure, but ultimately their processing is inherently mathematical and structured. It may mimic ethics, but it technically won't be ethics. Assuming that diligent programming would be geared towards optimising the efficiency of the robots, you might eventually see them calculating that it's more efficient for humans to be doing slave labour, or it might be viewed as efficient to the sustenance of finite natural resources that humans simply be eliminated. If you think about it, human beings aren't exactly working together to do things the smart, efficient and diligent way. Too many cooks spoiling the broth and all that.
The real thing of it is steps of progression in AI complexity and effectiveness. Allowing robots to process information and in a certain sense, become 'self-aware' wouldn't be any huge issue initially. As the decades pass, if that self-awareness is constantly accumulating, eventually you're going to run into problems.
But at the end of the day, if the human race isn't destroyed by space debris, it's going to be destroyed by hubris. May not necessarily be robots, but whatever it is will have hubris at its roots.