CurrentC will not fail because of this, they will fail because they use QR codes. An old lady could probably write a check faster than most people can scan a QR code.
If I'm reading that diagram correctly, not only does the consumer have to scan a code, but the vendor then scans another code. So not one, but two QR code scans. The success is built in!
So sounds like one or the other.
But unrelated, the steps you quote: gawd dahyum. I'd rather stand behind the old lady writing a check, as mentioned by another commenter. Thing is, retailers have mostly solved the check-writing problem (but they still can't make Mr. OldSkool fill out everything but the amount before he gets to the front of the line). No more "I need check approval on aisle 3", no more 3 pieces of ID, just scan it and stick it in the drawer. It sounds like there's a strong possibility that they're bringing us back to the time when those behind get to roll their eyes and sigh as we all wait for the person at the front because "oh, gawd, they're using that thing".
Except I doubt you'll ever see anyone actually use CurrentC.
Sure, they don't have to re-do the backend, but there's still figuring out how to actually effect the transaction between the customer and merchant. They're already talking a 2015 launch date for this, possibly longer (can't find the article, but people involved in the project are shooting to get it done within 2 years). Changing out the front end app and store support hardware/software would just make the thing more likely to fail.
It's technically possible, but one of those things that's deeply unlikely to happen, particularly given the lead that Apple (and, very likely, Android in the near future) and the credit card companies have in rolling out a solution.
EDIT: added a missing word