> Most of the technologically unemployed will wake up and do whatever AI tells them to do on a daily basis. Their lives will improve because AI is better than they are naturally at everything. This will lead to some weird outcomes. Especially if AI is not acting in the interest of each individual, but in the interests of the collective. This will cause AI to have to solve trolley problems.

Let's assume one second that AI becomes good enough to do that.

There's still a strong possibility that AI will be a tool acting in the interest of an elite, smaller (a few oligarchs, a single dictator) or larger (a country, a faction, a religion).