Suicide & Self-Harm

In AI safety, Suicide & Self-Harm is a content category used to identify potentially harmful material involving self-injury or suicide. Detection systems may look for explicit statements, encouragement, instructions, or other signals indicating a risk of self-harm so appropriate safety measures can be applied.

AI systems can encounter or generate sensitive self-harm content, creating significant safety risks for users who may be vulnerable. Identifying this content helps organizations apply safeguards, reduce harmful outputs, and support safer interactions while avoiding unnecessary intervention in benign discussions.