It's not hard to see that the scope of automation could increase in areas that would more drastically affect employment in the short to mid term than the other developments we've seen in the last 20 years. Either full long-haul automation or almost full automation with remote intervention could be a reality in the very near future, but almost certainly will be in 10 years. Call centers will be almost completely automated. Factories are automated to a much higher degree than they have been in the past. Every industry is solving problems with software that increase efficiency of workers (ie, fewer workers needed) and that trend is accelerating.
We can always say "this time isn't different". History tends to repeat itself. My point is: the trend is already there to see and the evidence suggests that it may accelerate very soon.
So far, there hasn't been problems with mass unemployment. But looking into the far future, when robots are advanced enough to do the basic labor of growing food, building homes, doctoring, shipping stuff around, etc. It's not super clear how much you'd really need human labor at that point, which could easily mean mass unemployment.
Of course, with the right societal design, mass unemployment could be a perfectly fine thing.
We don't have to guess how it plays out.