We'd need some sort of countermeasures to a robotic state. Something like a galactic - wide EMP that destroys all robots, but also cuts research & production by like 80%?
A galactic-sized EMP would be practically impossible due to myriad reasons, a handful being:
- You'd essentially need a humongous, galaxy-threatening WMD to produce it. The political consequences of that couldn't be overstated.
- Something strong enough to destroy advanced robotics would wipe out any and all post-industrial civilization in its path.
- I don't believe electromagnetic pulses travel faster than light, which means it'd take thousands of years for the pulse to propagate from ground zero to every corner of the galaxy. If it were multiple detonations, there would have to be thousands and thousands of them throughout the entire galaxy, and even then it'd take years for the pulse to reach many areas.
- I'm not sure, but I don't think EMPs just propagate indefinitely, without ever losing strength.
Anyway, I haven't read much on the subject, but it'd seem AI rebellions in Stellaris would solely depend on the amount of robotic POPs throughout the galaxy? Seems rather plain and unlikely, more mystical than plausible.
Robots from different empires wouldn't be networked other than via Spacebook or something (if that), and a galaxy-wide uprising seems far-fetched. I suppose there is a way, but is less a conscious rebellion and more like widespread corruption: a major AI entity could become self-aware and murderous, spread a virus and progressively subvert fellow robots, which would propagate the infection and establish communication with robots from other empires.
That sounds more plausible.