Preprint warns of catastrophic AI risks if no action is taken within five years

Nowhere is it discussing the risk of omnipresent surveillance and totalitarian censorship