The point is to raise the level of the bottom to something people can live on safely and with some level of dignity.
And with UBI in particular, it's to shift the impetus of your efforts from basic survival to improving your lifestyle from the lowest common denominator. People make shitty decisions with a gun to their head--with some room to experiment they'll usually be more successful.
Further, they might find other ways to individually create societal value (art, science, invention, philosophy) than just simply making their employer money. The only way you can do that right now is if you're already well-off or if some committee somewhere agrees what you're doing is valuable and subsidizes it. That's not a great way to innovate.
And they might not create value. Some people will sit at the bottom. But that needs to be OK--they do that now anyway while being a distinct drain on a society that wasn't built to accommodate that basic human nature. Nothing is perfect.
Most people work for their own self-validation if you give them a chance, though. Think we've all had that experience during time off where you're eventually compelled to do something productive just so you can know you did something.
So while some people might have their standard of living brought up, many (possible many more than would be brought up) will be forced down with the inability to pull themselves back up because of lack of jobs.
The best people, at least in industries where it fits, will move on to consult on the proper operation of the automation and others will become artisans doing highly crafted versions of what otherwise is mass-automated.
And some might move to other localities where the automation is too expensive to create or support. Automation has a cost relative to the market in which it's created and requires highly skilled people to create and maintain, whereas labor can be paid on a sliding scale based on their local cost of living and usually cost less than automation maintenance. That scales to low budgets better, especially if you don't produce quite so much of something as to hit the break-even point for automation.
UBI, in this context, is a mitigating strategy to keep the rest from becoming welfare recipients while they retrain or retire early. That would be a hell of a lot worse stigma than simply having a lower standing of living.
Automation mostly doesn't "eliminate jobs" (numerically; it certainly might eliminate job categories), it reduces the market clearing price of labor by providing a competing alternative.
Among the features of a mature UBI is replacing the combination of minimum wage and means-tested benefit programs as a means of ensuring that the "fully employed" and the less-fully-employed, respectively, are adequately supported. By displacing minimum wage, it actually reduces the degree to which reducing the market value of labor results in lost jobs.
Now, it doesn't eliminate income or wealth inequality, which is essential to maintain some incentive to engage productively with the economy for those who can, but its not a matter of nobody working, its a matter of mitigating the cost of a narrower class of people having high-value labor while the labor of the masses is increasingly low-value given the alternatives available.
Let's say we institute a UBI and 50% of jobs are eliminated through automation. Everyone has a basic standard of living, but those who still have jobs (who designs the robot?) have a much better standard of living (they would have to or why work?). Now you have a huge amount of income inequality and 50% of the population can't do a dam thing about their situation because there are no jobs.
The only way to fix that is to redistribute so much wealth that the incentive to work is basically eliminated. I think that's been tried before and it failed horribly.