Follow
Mastodon feature request
An AI-powered pre-flagging feature that can highlight to admins content that might warrant some sort of ban, based on previously flagged content. An opt-in feature could hide the content automatically until a mod reviews it. This would only work for text, not media.
I feel like this would be useful for big instances that don't have many mods.
Mastodon feature request
@realcaseyrollins you could probably implement this as a mastodon client bot without messing with the server code itself.
Have it monitor the local timeline and pipe it to spamassassin. If a message scores too high (i.e. likely unwanted content), the bot can fire off a report. If you gave it mod privs, it could also silence the user or post pending a human mod looking at it.