Even if of questionable quality it can be used by mappers to speed up their work, that’s the same principle as rapid or mapillary, they help us speed up the work, or at least I used them this way (but I don’t think everyone else would like to map every building in a city by hand), and if it’s not the best quality we can modify the data and tweak the quality by hand, I don’t import dataset on rapid as they are if they don’t conform.