AI-Powered Scam Detection: Community Discussion
Summary
The CommunityOne Builders Help server recently hosted a discussion regarding the feasibility of using AI to detect scam patterns in chat. The conversation explored the limitations of regex-based filtering and the high computational costs associated with running AI on every message. This dialogue highlights the ongoing challenges developers face when building automated moderation tools for Discord communities.
I would love to see a feature where the bot looks over the chat and detects odd/scammy behaviour, for example you make it search for sentences that look like "Join here for support" and it automatically detects it and notifies the channel/deletes a message. This would be super helpful for moderation and a great way to filter scammers
oh yeah.. will keep that in mind!... it's hard to do it based on regex though....
I thought the AI could detect it by itself instead of a regex
thatd require us to run AI on all the messages.. which can cause a fortune lol
oh fair enough lol
