Just to throw my 2¢ in, I too have fantasized about bot/human differentiation in NodeInfo, especially in the context of https://encyclia.pub eventually opening up to all ORCID users. Whenever that happens, its account stats might abruptly jump from 16 to 15,000,000 if I report all reachable accounts.
Encyclia happens to be a bot-only software and could be manually excluded from stats, but clear labeling and aggregation would be even better.