Another way in which MWI is falsifiable is that it would be proven wrong if those Chinese physicists were able to show a speed limit on the spooky action at a distance. MWI says that there is no such speed limit, since there is no spooky action at a distance.

Additionally, there are collapse theories that are much easier to test than consciousness-based ones. By "much easier", I still mean fantastically difficult, but should the day come that one of these collapse theories is proven correct, then MWI will clearly have been falsified.

You can set up complicated and impractical experiments (for us, but perhaps not for inconceivably advanced aliens) that can experimentally distinguish between MWI and other interpretations of QM. (With the exception of Bohm.) Do you really want to know all the gory details? They are rather obscure.

The executive summary is that it is in theory possible to determine whether a system is in a specific superposition of states, if you know exactly what superposition of states it might be in. So, let's say you wish to see if MWI is true vs a version of Copenhagen that relies on a conscious entity experiencing the results of a measurement in order to cause the wave collapse. If you have an accurate enough model down the very last particle of the entity's brain, and everything else the entity would have to interact with for it to perform an experiment, you can in theory then experiment on the entity to determine if it (plus everything it interacted with) is in the specific superposition of states that MWI would predict.

As I said, not very practical.

Well, David Deutsch argues that it might be practical when the day comes that we have AI's living in quantum computers. (And his experiment is also a bit different from what I described, and a bit more conceivable for us to do with technology we might someday have.) A problem there, is that naysayers will no doubt claim that the AI's aren't conscious.