Our Approach to Content Moderation

Moderation across Discord

All users have the ability to report behavior to Discord. User reports are processed by our Safety team for violative behavior so we can take enforcement actions where appropriate.

Discord also works hard to proactively identify harmful content and conduct so we can remove it and therefore expose fewer users to it. We work to prioritize the highest harm issues such as child safety or violent extremism.

We use a variety of techniques and technology to identify this behavior, including:

• Image hashing and machine-learning powered technologies that help us identify known and unknown child sexual abuse material. We report child sexual abuse and other related content and perpetrators to the National Center for Missing & Exploited Children (NCMEC), which works with law enforcement to take appropriate action. 
• Machine learning models that use metadata and network patterns to identify bad actors or spaces with harmful content and activity.
• Human-led investigations based on flags from our detection tools or reports.
• Automated moderation (AutoMod), a server-level feature that empowers community moderation through features like keyword and spam filters that can automatically trigger moderation actions.

In larger communities, we may use automated means to proactively identify harmful content on the platform and enforce our Terms of Service and Community Guidelines. 

If we identify a violation, we then determine the appropriate intervention based on the content or conduct and take action such as removing the content or suspending a user account. The result of our actions are conveyed to users through our Warning System.