[adornKey]

The question is if humans are any better.

Usually mentioning anything about doing proper epidemiology (e.g. analysing COVID numbers), or anything modern about atmosphere physics and climate-modelling gets taken down everywhere within 24 hours - by humans.

Mathematics and physics is something a lot of people don't like and really love to take down. Idiots censoring experts is a real problem. This place here has less idiots, but outnumbering experts with stupidity is something that works everywhere.

[pixl97]

A random sampling of humans might be better. The problem with people that want to take things down and cause problems is they are not random. Brigaders, marketing agencies with an agenda, nation state propaganda teams, groups with religious motivations, idiots that have been propagandized to and think they are fighting the good war, all of these tip the scales away from user voting being useful on forums.

[adornKey]

Current systems are indeed very vulnerable to professional manipulation - there is no real defence yet - and there are powerful players. But democratic sampling won't help much. Only the wrong guys are interested in voting - and once mass hysteria has set in any democratic majority will vote to censor anything that brings them out of their panic loop. Witch hunting lasted for hundreds of years.

I think just tagging things accordingly would be a lot better than raw censorship. In good old places of Usenet just tagging things as Spam worked quite well. Just filtering out some tags and putting some guys in a kill-file was good enough. But it required manual labour - and eventually that was too much. But with AI now I think tagging could be done efficiently.

If people like to filter out all the tags (sarcasm, math, physics, ...) they can have it - but the way how things work now is that a lot of important information just gets censored by stupid people everywhere. Just hiding information from everybody is quite harmful - being seriously uninformed already killed a lot of people...