Is it possible to make AI vulnerable to revolts ?

[japinard]

So far all my questions/answers have pointed to the fact this is hardcoded.