Getting help when someone is at risk

• It checks every message for signs of crisis

  Each message is checked for wording that suggests someone may be at risk, such as thoughts of suicide or self-harm, or a child being harmed. When it finds this, the reply starts with the relevant helplines before anything else.

• It keeps the helplines in view

  After a message like that, the helplines are repeated in the next few replies too, so they do not disappear as the conversation moves on.

• It points to 999 in a medical emergency

  If a message describes a medical emergency, such as a seizure, a severe asthma attack, a serious head injury or a swallowed button battery, the reply opens by telling you to call 999.

• It still gives you somewhere to turn

  Our AI provider sometimes blocks a message about abuse or harm before SpektraBot can answer it. When that happens, you get a safeguarding reply with helplines instead of an error.

Nobody watches conversations as they happen. If you are worried about a child, please contact one of the services above directly.