Safety hub

Moderation & Communications Review

Version 1.0 · Effective 2026-05-22

When communication features are available on CombatScore, we use AI to automatically review certain conversations for safety. This page explains what that means.

Who we monitor automatically

  • Any conversation involving a minor athlete (under 18). Mandatory. Cannot be opted out of.
  • Coach ↔ athlete conversations where a power dynamic exists. Mandatory.
  • Any conversation that has been reported by a user.
  • New coach accounts in their first 30 days (heightened scrutiny period).

What we look for

  • Grooming patterns
  • Off-platform contact attempts
  • Sexual content or innuendo
  • Self-harm references
  • Substance references
  • Age-inappropriate topics in conversations involving minors
  • Gift offers from adults to minors
  • Attempts to arrange physical meetings outside of the gym context

What happens when something is flagged

Flags are categorized by AI confidence:

  • Low confidence: logged only.
  • Moderate confidence: queued for human review.
  • High confidence: thread auto-locked, gym owner and parent notified, safety team paged.

A human reviewer can clear false positives at any time.

What we don't do

  • We don't share message content with third parties unless required by valid legal process.
  • We don't sell or monetize message content.
  • We don't use messages to train external AI models.
  • We don't analyze adult-to-adult conversations without a report or report-equivalent trigger.

Retention

  • Unflagged messages: retained 90 days, then deleted.
  • Flagged messages: retained 7 years to support investigations and litigation.

Your rights