A concern about bulk changes to Citation Quality is that it is governed by the fact for which it is evidence. So, not all citations of the same source should necessarily be of the same quality.
Just a follow-on Tom's comment: I don't use the Citation Quality field provided by RM, although I know I should. Part of the reason is that it just sort of sits there and does nothing. It doesn't show up in any reports and it's not really "actionable", for example you cannot test it and take action based on the results of the test, you cannot you search for it and make groups. etc. You can see it yourself within your RM database, but that's it and it's pretty hard to see.
And as Tom pointed out, the same citation may have different quality depending on the fact for which it is evidence. So if you memorize and paste a citation, you may well create an incorrect Citation Quality. It might seem that my use of extremely split sources would aggravate this problem in my own use of RM because I put all the data I can in RM's Master Source rather than in the Source Details, and Citation Quality is stored in Source Details, and correctly so. But being an extreme source splitter doesn't really matter, because a source lumper will have the same problem after memorizing and pasting citations.
I think this issue has implications if ever RM were to implement something like "shared citations" such as are supported by FTM. With RM, if you memorize and paste a citation many times, it's hard to find them all and make changes or corrections to each one of them. With something like FTM's shared or citations or with my own extremely split sources, you can make changes or corrections only one time and the changes or corrections immediately apply to all occurrences. It's not clear to me how to maintain the ability to make changes or corrections to shared citations only one time while at the same time maintaining the notion that Source Quality is governed by the fact for which the source is evidence.