I think this is an oversimplification, because in practice, there will be a gradient of people around the inflection point who would be willing to take less in order to stop working, and that gradient will vary by the individual and by the job.
Additionally, I can also imagine politicians promising increases to the UBI without concomitant tax increases, funded by bonds or other kind of financial mechanisms for monetary injection whose long-term negative externalities would be easy to obfuscate from voters.
> Imagine a world with widespread automation and general or close-to-general AI without a UBI.
This is the key point. UBI is (in my current opinion) a necessary tool for smoothing an economic transition to a highly-automated level. However, I am concerned that it might induce other economic instabilities such as in tax rates, voluntary or involuntary unemployment, underemployment, inflation, borrowing, minimum wage, and so on -- these instabilities could then spread to social or political disruption that would undermine progress towards the transition that UBI was originally meant to support.
Since the size of the UBI would roughly correlate with the size of its influence economically/socially/politically, then as the UBI increases, the variance of these potential instabilities would also increase. As such, increases to the UBI should be carefully controlled, for example by gating them to developmental milestones. Of course, that opens a whole can of worms on who sets the milestones and so on, so I suppose it's really just shrugging and hand-wringing all the way down.