You're just shifting the trust to the "data bank". What if it gets hacked, or ignores its ToS, or some rogue employee with database access copies your PII, or... ?
Trust is dead; it doesn't scale. End of story. I don't like the idea of a "data bank" or "data brokerage" any more than I like the idea of facebook; either way, someone else holds the keys to my castle. But third-party servers aren't going anywhere, so we have to find a way forward.
I've come to the conclusion that the only possible way for someone (or something) to maintain control over data is to encrypt it at all times when not in use. It's exactly the idea of a "data bank" like you're describing, but you become the bank. One of the key struggles with this approach is that it requires a tremendous amount of client-side code; calling it a paradigm shift is a profound understatement. It doesn't matter how snazzy your encrypted-dist-web product is unless you present a compelling economic reason for companies to switch to it, to justify the extreme expenditure of capital necessitated by the technical switchover.
Even here, though, once you share something with eg. Facebook, you can only hope they don't fuck it up. That's just the way the world (unavoidably) works; once you share something with someone, they have every capability to do whatever they please with it. You have only the social expectation that they don't, and if they violate that expectation, your only recourse is to stop communicating with them. But I do think, in a world where we actually have autonomy over our data (partly, again, because it necessitates client-side code), it is possible to make the consequences of data misuse so disastrous for a company that it stops being economically viable for them to do so.
The key thing we've lost is that agency to make decisions about data. I've no problem with informed, consensual sharing, but in today's world re: data, "informed consent" is nonexistant.