They've said they're working with an external IRB. Not telling them would be a massive issue for informed consent, and at least if I was sitting on said IRB would generate a pretty firm "Um...no."
A big part of this experiment is that we want to see how (and if) basic income will change people's behavior. To be able to do this, the income stream has to be reliable. That's why the idea is often called Basic Income Guarantee or Unconditional Basic Income.
If the income stream could just stop at any time, you couldn't measure its psychological impact as effectively.
Why not give the subjects a basic income for life? That will mirror the potential reality of basic income. You could still only run the study for a year (or five years) but you'd have to pay out for the rest of their lives if you want the subjects to behave in a realistic way.