Response to all of these: It depends. Sometimes.
I identify swans in my area, and there are two similar species. So I went through all of the observations and found around 5% false IDs. I also added some third or forth IDs to correctly identified ones to precisely make it harder to be changed. Because I’m an expert in these species, and I can ensure RG observation of the species stay RG and IDed as that species. There are not many new swan species popping into existence. But if there will be a cryptic species, all RG observations will be in need of moving to a higher rank anyway, and it should be handled as a taxon split and not by individually challeging previously correct ID.