There also numerous examples of failed mass uprisings. If you can throw enough robots at the problem, the 0.1% would have nothing to fear.
When high automation comes, it's not going to be only the rich that own robots. As an analogy, take smartphones - there are 5 billions of them in use in 2017. Almost everyone can have the best phone today, or one that was as good as the best from a couple of years ago.
There are going to be cheap and useful robots for everyone, 3d-printed, open sourced designs running on open source AI, and capable of assembling more robots. What goes into a robot? Motors, gears, sensors, batteries and computer chips? Those are all commodity everywhere on earth. The moment we can assemble them into working useful robots, we will see an explosion of do-it-yourself automation everywhere.
With self printing and self assembly we can make a software based factory that can "compile" itself and anything we desire. We will have built the first "material compiler" that can "bootstrap itself" by creating a replica. Then everything will cost only as much as raw materials.
If these robots can move, not even “requires physical access” will save you from the Zombie Robot Apocalypse.
> Then everything will cost only as much as raw materials.
Seems like one would want to own the raw materials, then. But if the evil Dr. Robotico's robots can kill the owners and take their quarries, then what? Does he rule the whole world? Are we in a race to robot armageddon?
Man, that Paperclips game is starting to seem more and more prescient.
As a very quick thumbnail, I'm trying to blend Integrated Information Theory, Selectorate Theory, the eusocial dynamics of colonial animals (including ourselves, per Wilson), and some probably-wrong ideas about meiosis (cell division) to come up with a biologically-rooted model of the collective action problem. Not being a professional scholar, there's little or no chance I'll ever weld these into a general theory of human social behavior, nor do I want to; I hope to intuit a good-enough grasp to deploy it within polities and rely on its fundamental harmony with nature as the least destructive method of manufacturing large-scale social change. It's probably a doomed endeavor but I'd rather fail at a hard problem :)