AI Safety Tools
How AI helps us spot harm faster and keep members safe.
AI does not replace our team, but it helps us look at more posts, faster. Here is what it does, what it does not do, and how we keep it accountable.
What AI Does
AI sorts new posts by risk so the highest-risk content reaches a human reviewer first. It also flags duplicate spam, suspicious account creation, and known abuse patterns.
What AI Does Not Do
AI never bans accounts on its own. Every enforcement action is approved by a person. AI is a triage tool, not a judge.
Keeping AI Honest
We test our models for bias each quarter. If a model treats one group worse than another, we retrain or shelve it. We publish results in our transparency report.
Member Controls
You can ask us to stop using AI for actions that affect your account. Reach out via Settings → Privacy → AI Use. We will switch your account to human-only review within seven days.
Quick tips
- •Read our AI methodology page for details.
- •Reach out if you suspect a biased flag.
- •Use the appeal process for any AI-flagged action.
- •Stay current via our policy mailing list.
Related articles
Keep exploring our trust and safety library
Threat Detection
Systems that watch for unusual activity and stop attacks early.
Read moreTransparency Report
Twice-a-year numbers on requests, actions, and outcomes.
Read moreWas this article helpful?
We update Trust & Safety articles based on member feedback.