They're not exactly sure when this will happen, but one thing they know for sure is the Gulf Stream has definitely lost its stability and is in the process of slowing. A collapse will mean the end of the world as we know it, and it is expected to happen quickly.
Lots of attitude there, but no common sense. 'Sure when it will happen' is like asking for infinite precision; once you get upper and lower limits (one week, one millenium) you already know 'when'. As for disprovable, that's what you want in an hypothesis, it isn't relevant when a model makes the useful prediction that justifies its existence.
A lake in WA state was getting more sewage than was healthy, and its oxygen-depletion death was predicted, SO our local governments formed a coalition to reroute that waste, and clean up the lake. It worked. That doesn't mean the prediction wasn't scientific, because TESTING a prediction is uninteresting, unless it tests a (questionable, uncertain, untested, controversial) hypothesis.
Scientific predictions are sometimes warnings of things we can prevent.
Traditionally the computational models have underestimated the AMOC current effects by a factor of 10 as compared to ocean buoy measurement data. The tipping point has as already been exceeded for some time now, and the they're really watching the transition to the slower/ weaker state.
Here is a little more on the same paper:
Title should be ":collapsing" because that what it is. And now that it's started, it's irreversible.
Here's some of what they have in place to monitor this complicated mess of interfering factors.