Here's an interesting interview with Dr. John Hopcroft about the future of job automation and how we could manage its ill effects on the human population:
I agree with his point that, just because humans have been able to climb up the skills ladder in the past faster than machines could automate old jobs, doesn't mean we will be able to do that forever. Past trends don't continue indefinitely, and there's no reason why we couldn't get into a situation where machines took over 1 million human jobs in a given year, but only 900,000 new jobs were created during that same period. In fact, I think that's what will happen.
While, at first glace, it might sound great not having to work, the reality of the situation will be far more dour, at least in the short-to-medium run. As the pandemic-induced unemployment spike in the U.S., coupled with the wave of civil unrest shows, idleness and inadequate government benefits lead to widespread discontent and political instability.
Hopcroft suggests that we could ease the pain by effectively spreading the remaining jobs out among a larger number of people as time passes. Higher education could be drawn out for a longer length of time (this probably means we subsidize college and accept that no one will get jobs until they've finished a masters degree, then later a Ph.D., and maybe later on some higher level of accreditation), people with jobs could be given more mandatory vacation time, and retirement ages could be lowered.
This is surely a conservative's worst nightmare, but from a logical and economic standpoint, it might be the cheapest way and perhaps the only way to maintain social and political stability. I think these reforms could be thought of as an across-the-board expansion to the existing welfare state, and I believe some centralized, homogeneous countries with culturally obedient people will pull it off this century. However, as an American, I have no faith in our ability to do the same, unless it comes very belatedly and after a prolonged period of internal acrimony.
Let me also say that the across-the-board expansion of the existing welfare state, a little bit each year, is much more realistic and more palatable than the institution of a UBI created from whole cloth. For obvious reasons, it would be easier to build upon government programs that are already familiar and accepted by average people than to create an entirely new and massive one. A constellation of many such programs, like more college tuition support, expanded access to state healthcare and housing vouchers, and mandatory paid vacation days, would also effectively accomplish the same goals as handing out cash.