Someone should set up a third-party tool that mastodon/fediverse users can report abuse to. It should be staffed by people who are skilled at evaluating this stuff and who are paid to do it.
Recommendations/instructions to block instances/individuals could then be sent out to instance admins or even individual users (in an automated way), either as a collective or by subscription.
Or, rather, is anyone doing this already? #federatemoderation
@youronlyone In some ways, yes. But we need to remember that people almost always have some level of reputation. What makes the difference is that reputation isn't a linear number value, but a wide set of different scales (per topic etc.).
Problem with digital identity reputation is that it gets more difficult over the time when amount of accounts grows. Even more in a federated system where people can setup new instances & servers in a few hours.
@Robotbeat yup! To a point. In some ways, this could be construed as inventing a whole new set of de-facto laws around engagement.
On the other hand, individuals don't owe anyone anything. I could block everyone who uses the word "orange" and the only one that would suffer is me. 😅
On Facebook and Twitter, the ability to appeal mattered a lot more, because there were no other communities to go to. That's not true here.
If ordinary users are going to be “rated” by various organisations and instance moderators, then there should also be “reputation scores” for those organisations and instance moderators (and their instance itself).
It should go both ways.
Is the server, and/or organisation, abiding by The Santa Clara Principles or not?
Users can then be more informed which server they want to join. Same with admins trusting third-party “signals” from these organisations.
@blaine an outside moderation tool will, I think?, always have less granularity than the instances' preferences granularities. As in, you could have a vendor provide a few specialized views into the fedi, but any shared blocklist is going to have some amount of misalignment with any specific instance's preferences.
I wonder if there's a way to share moderation decisions by weighting the influence of neighboring nodes in an affinity graph. (this isn't the same as the federation graph, ragelove might be willing to federate with fosstodon but have no trust in their moderation decisions.) If aleph high-trusts ragelove, and ragelove defeds naziparty, how can aleph benefit from that decision automatically? Maybe if multiple high-trust peers send along the same moderation recommendation, or several medium-trust peers, etc.
(This sounds kinda like "whuffie" in an early Cory Doctorow novel. I may be cyberpunkpoisoned by reading it during an impressionable period.)
@gsora e.g. if the hosting instance (e.g. p o a . x t) doesn't block the user, then either the network needs to defederate from the server, OR communicate that the user should be blocked everywhere.
But actually this speaks to the need for formal organizations to help with this. I'm not an expert, and don't really know the answers here. It's important that we delegate and help each-other where we can.
That is already happening and has been abused before, and is continually being abused today.
Bad actors create fake accounts, pretend to be real accounts, then launch their server-disinformation campaign. "Server-disinformation", meaning, they will post content most server admins hate, then they will report those accounts in various blocklists and servers, with claims such as, "this server is so and so".
Server admins blocks the entire server w/o investigation bec. of 1 user.
@youronlyone @gsora absolutely. This is a really good example of why we need sophisticated folks who can evaluate these sorts of attacks and make the right call.
It's unreasonable to expect admins of small instances to be able to do this themselves.
@peepstein @timbray @blaine I agree. That we're even discussing mandatory centralized authority structures and imposing costs just to participate indicates that we're missing the point. The value of ActivityPub is in its permissionless and decentralized nature. We should resist the impulse to try to build authority in as a default.
I'm all for secondary services to allow people to pool resources around categorization. But participation shouldn't be contingent on submitting to such a service.
Um, have a look at this thread, it may help explain why we're sounding obsessive about our defensive alignment: https://twitter.com/rahaeli/status/1594724708309553152
@timbray @peepstein @blaine Thanks for that; it does help contextualize the conversation for me.
I don't disagree with any of that. If you run a service for the public, you need controls in place to minimize abuse.
I just starting admining a small server and one thing that was super annoying was trying to copy over block lists from other instances that I trust.
Along those lines maybe the reason for blocking an instances becomes more important... community standards vs. hate/etc.
Is it just me or is the user experience around managing that terrible?
blaine
Daniel Schildt
MoveFastAndFixThings
Paul d'Aoust
♾️ Yuki (스노 雪亮) 🐬 🦣
Andreas Rilinger
Robotbeat
external quantum efficiency
Aleatha
skry
Matt Secoske
Tex
Leigh Honeywell
Aaron Huslage
Going T. Maine
Ms. Penny Oaken
Gianguido
Bret Mogilefsky
f00f/eris/continuum/etc
Tim Bray
GenghisKen Coar
Benjamin Richter
Ed
Justin Thomas 🛡
Mark Madsen
Sasha Costanza-Chock
nev