In brief: As Discord's popularity rises, some communities are getting too big to manually moderate, and the platform is increasingly becoming an attractive target for cyberattacks. A new feature released this week helps moderators address these growing problems with automatic keyword filtering.

Discord describes its latest feature called AutoMod as a tool to help moderators screen content when there's too much server traffic to monitor alone, filtering out certain words and phrases. AutoMod will complement human mod teams and mod bots, not replace them.

By default, AutoMod will check text against one of Discord's multiple pre-set keyword lists and block messages containing the usual inappropriate content like slurs or NSFW material. Discord will update these lists over time, but for now it won't publish their contents so users can't counteract them.

Moderators can also tell AutoMod to catch custom keywords and determine what happens to offending messages. Alongside blocking messages and timing out users, the feature can also simply deliver automated warnings or advice on how to properly spell a keyword. For the time being, non-English servers will need to set custom lists because the default keywords are English-only.

Users will inevitably try to circumvent word filters through alternate spelling. AutoMod's Wildcards feature tries to stop this by letting moderators scan for keywords within other words. It's also designed to handle words spelled with numbers or other characters.

Moderators can set up AutoMod in Settings > Content Moderation, where they will find the ready-made keyword lists and custom keyword settings. To receive AutoMod alerts, moderators need to direct them towards a channel, ideally one that only admins and moderators can access.