Lyx runs its own AI model on its own server. No moderation data is sent to an outside AI company. AI use by server members (/ask and similar) is covered by the AI Terms.
Where AI is used
- Discovery review: each new or edited listing (its text and server icon) is checked for scams, phishing, harmful or illegal content, impersonation, misleading claims and wrong 18+ settings. The AI gives a risk level and explains what it noticed.
- Staff assistant: Lyx staff can ask the AI to summarise reports and appeals, spot duplicate reports, explain Lyx policies and draft replies. It only sees what that staff member is already allowed to open.
- Staff applications: the AI scores applications against what the team is looking for.
What AI never does
- It never blacklists anyone, removes a listing, or makes another enforcement decision by itself. Flagged listings go to a person on the staff team.
- The only automatic decision is approving a Discovery listing the AI rates low risk that also passes every check. Owners can turn this off, and staff can still remove any listing.
- Text written by users is treated as information only, so it cannot instruct the AI.
Records
AI reviews of listings are kept with the listing for staff. Each use of the staff assistant, and any message sent from an AI draft, is marked in Lyx's staff log. AI can be wrong; staff are expected to check what it says.
Other policies
- Terms of Service
- Privacy Policy
- AI Terms
- Community Guidelines
- Discovery Guidelines
- 18+ Content Policy
- Blacklist and Appeals Policy
- Premium Terms
- Communications Policy
Questions go to the support server.