hardy tangle
Hey guys! Thought you all might be interested in seeing this automoderation module we've launched on the Leonardo.Ai Discord server, in our Discord bot Verrocchio. It reads messages of people who haven't talked on the server much yet to see if they break any rules. It then decides what action to take based on the rule. The "Other" category only silently logs so that we're aware a member may have broken a rule, whereas the "Hate" category deletes the message and times the user out. All categories except "Other" result in a warning being issued by Verrocchio to the user as well.
On top of this, Verrocchio also scans usernames for potential guideline breaches, and can change them if deemed appropriate. It then finds an alternative that keeps the user recognizable while no longer breaking rules. There's also a "Scam" category which on top of timing out and deleting message also adds the message contents to a list of known scams, meaning next time that same message is sent it is instantly flagged as a scam without even needing to go past GPT.
Finally, our users can also talk to Verrocchio in the way they would to a human moderator, both in the server and in DMs. Verrocchio can answer many questions thanks to its extensive database (and does so automatically when it finds a question it can answer), and also contact moderators when necessary. This means users can report certain behaviour to Verrocchio, after which Verrocchio will make an open issue about it which the moderators can investigate.
When a breach of our guidelines is reported, moderators are presented with various buttons they can use to act on it. If they select Ban, Verrocchio will generate a ban message for the user as well and DM it to them, along with the option to appeal.
All in all, this is an extremely advanced addition to our bot that significantly enhances the server! Would love to hear your thoughts on it!
worldly comet