The challenge with moderation in digital communication is that disrupting communication often scales better than individual blocking.
In the Freenet Project (where centralized moderation simply is no option) the answer was to propagate blocking between users in a transparent way. That way blocking disruptors scales better than disrupting. For more info see: The Freenet Web of Trust keeps communication friendly with actual anonymity
If you want to try this, there are three steps:
- First the current state in Freenet: OadSFfF-version1.2-non-print-edition.pdf
- Then the optimizations needed so this scales to arbitrary size: deterministic-load-decentralized-spam-filter
- Finally some starting code of a more generic prototype for faster testing: hg.sr.ht/~arnebab/wispwot
Also here’s some data of the real world communication structure within Freenet that should be useful if you want to test algorithms: The_Freenet_social_trust_graph_extracted_from_the_Web_of_Trust
(the prototype implementation is built in a way that would be suitable for federation, because it can work with a shared database that only has different entry points to get your personal view of the trust-graph. It is far from finished, though)
I hope this helps you tackle these issues, because global-scale moderation without centralized control is one of the huge tasks ahead of us — a task that was mostly ignored in the Clearnet (there were underpaid moderators to burn out after all) but tackled within the Freenet Project more than a decade ago.
I wrote this as a reply to an article about decentralized moderation.