AI Glossary / Content moderation

🛡️ Safety and quality

Moderation in AI context is automatic content checking for prohibited categories: violence, sexual content, hate speech, self-harm. Filters apply both to what users input and what the model generates - and moderation systems are themselves AI models. For app creators, providers offer separate moderation tools worth placing before and after the main model. For users, moderation can be frustrating with false positives - it's responsible for refusals on innocent but ambiguous-sounding requests.