I tried to explain to T-compiler, some time ago, that opaque number sequences ("issue numbers") are not very informative about what is behind that pointer, and pretending they are is harmful. People could provide, instead, actual reasons for things, which do not require dereferencing random pointers, which thrashes cache.
I feel that has been demonstrated here.
But yes, I think probably we need to rethink the label-assigning approach. Teams could take it on themselves to submit an actual report of what they'd like to see in the release notes, for instance.