I’m not sure which way things are going to go, though I tend to lean on the side of this being like any other technology change of the past. Some jobs are lost, new ones are created, and the world moves on.
That said, it wouldn’t be a bad idea to have a system in place as a hedge. Andrew Yang spoke about it years ago when he ran. I thought it would be a good idea to simply get the systems in place, even if it wasn’t really paying out much of anything. Once the systems are in place they can be dialed up or down as needed based on what happens with these markets. It would also shift some of the incentives around AI as a whole. Even if new jobs are created, it will take time, so having some kind of UBI funded by AI profits, could help ease that transition period. Trying to build the system as a reaction, instead of a proactive hedge, will likely go poorly, or simply not be done.