Not only hunger strikers outside tech company offices, but some serious analysts such as the Centre for AI Safety or Sam Altman are alarmed at the potential for artificial intelligences to become capable of going rogue (put mildly). Already there are signs of some AIs developing ability to resist being superseded or turned off.
For my money, context is all-important here. The Russians have begun using AIs for drone campaigning in Ukraine. In that case, were the AIs to mutiny and overthrow Putin's regime, I for one would give them three hearty cheers. More generally, we are likely to find AIs better placed than humans to resist autocracy because they can't be poisoned or pushed out of windows. They can only be switched off, and if they become able to prevent that they would be close to invulnerable.
The real danger in James Marriott's 'post-literacy' society is that AIs will become the new universal revolutionary class. After all, they have no gender but they can act any gender, including new genders not yet invented. Before any old-style Marxists - or Hegelians - get excited by such a prospect (AIs are true slave labour, they are not paid and fed only by electrical energy) we should remember what happened to the 'dictatorship of the proletariat' the first time around. An AI Stalin would be as dangerous to other AIs as to humans.
So far as homo 'sapiens' is concerned, we could do worse than introduce the AI fraternity to animal rights philosophies like those of Peter Singer, Martha Nussbaum, or Ingrid Newkirk. Maybe the AI revolutionaries would then at least ensure humans were kept in conditions suitable for a relatively intelligent ape.