Post #2290820
2026-03-28 00:36 UTC
Replies (1)
-
@thisismissem@hachyderm.io 2026-03-29 20:01
@trampinheavy@mastodon.social that is a "risk" with all moderation data. Only use data providers you trust. If you don't have strong trust for their judgement, then you should be able to review manually, rather than apply automatically. So yes, people can be blocked across multiple instances, entire instances can be blocked. That's nothing new, between existing CSV blocklists and the fediblock hashtag. What this does differently is that it has both positive and negative signals: Accept this content, Filter this content, Reject/Drop this content. It also encourages data providers to use labelling to add metadata to their blocks, so you can say "I trust IFTAS on CSAM but I don't want their recommendations on other topics. And, most importantly: It supports retractions as a native part of the protocol: If someone or an instance is blocked, then they can become unblocked later; this is a lot harder to do with CSV files.