I think it will be necessary in the medium to long term, but it will be very difficult to implement (and I don't mean for political reasons only).

Currently, we should intensively test all kinds of different UBI models to figure out what works and what doesn't. It is too early to roll out UBI yet, because for example

• "The machines" are not yet able to perform all tasks that humans wouldn't do in case of UBI
• We have just so much work to do to set the conditions for our future right. The biggest part of the world is technologically decades behind what is already possible and that also limits very much our potential to automate stuff (you can't use AI for pen-and-paper bureaucracy for example). We need all hands on deck for the transition, so to speak.

Concerning what works / doesn't work, we need to figure out a kind of UBI that

• Still ensures that tasks that need to be done get done / doesn't totally disincentivize "work" / economic activity
• doesn't lead to everyone just "lying flat" and civilization slowly rotting to death
• doesn't cause runaway inflation . I guess, we will need to find a way to make the amount of UBI dependent on productivity or sth like that.