Mass of people that will not revolt and storm the castles because they will have computer games, recreational drugs and not so bad life.
That will not care about voting so 1. would not happen and 2. would not end up in French style revolution as they wouldn’t mind state control.
And if AI is what the developers hope it to be, the fictional reference points would expand to anything from Manna to The Culture to I Have No Mouth.
I do not believe Approach #1 has ever been attempted without quite a bit of blood leading up to its attempt, and I think it has only been attempted a few times even so, approach #2 - oppression, with resulting bloodiness - seems to be the norm.
https://www.nybooks.com/online/2009/12/09/1989-poland-was-fi...
One reason for that is that when an occupying power leaves regime change must happen.
That happened for example in Romania in 1989, somehow it did not happen in Poland.
Look at Syria for how a similar pattern of events did lead to civil war. Ceaușescu just didn't represent a section of the country with an interest in fighting for his regime. The risk of instability when an authoritarian passes is if they have been holding together a country with significant ethnic/sectarian/political divisions like Yugoslavia.
https://en.wikipedia.org/wiki/Solon
Of course Oligarchs still took power afterwards from time to time - The Eupatridae (Aristocratic Families), then tyranny by Peisistratos.
> If you lock in wealth inequality at today's levels (or make them worse) and then use repressive means to prevent any form of push-back, the state becomes brittle.
A panopticon thus exists in phases. first phase (centralized) to buy time, stabilize a complex society, then to help forward a gracefull decay, finally (decentalized) to prevent the decayed and loopstuck society from further self damage and keep complexity recover vectors open.
The dream is a Afghanistan with phones that can bootstrap itself out of the failed state.
Historically speaking though, wealth inequality in developed nations is reaching a point at which, as often as not, revolutions happen. Sometimes they're bloody. Sometimes not. There are two approaches to avoid a revolution:
1. Back off from controlling government and let reform happen gradually. Your wealth will be lessened, but you'll probably remain one of the richest people alive.
2. Lock everything in with a totalitarian panopticon state that you control. AI surveillance offers a huge advantage over surveillance networks used by past totalitarian regimes. e.g. East Germany's Stasi showed the limitations of surveillance in an age where humans had to do it themselves. With AI to do it, you can avoid employing a large portion of the population to watch the rest, and you can keep the power of that surveillance network concentrated in just a few hands, preferably yours.
It's clear which approach is being attempted in the U.S.. I'd just point out that #2 is inherently high-risk. If you lock in wealth inequality at today's levels (or make them worse) and then use repressive means to prevent any form of push-back, the state becomes brittle. If it breaks, the result could be a bloodbath. Approach #1 is much safer. Yet, here we are.