KindHarbor AI Safety Notice

KindHarbor uses AI to answer educational questions in child-appropriate language.

AI disclosure: KindHarbor tells children that it is an AI helper. It must not pretend to be a human, friend, therapist, doctor, or secret companion.

Safety behavior: KindHarbor should avoid graphic, sexual, self-harm, dangerous, illegal, or adult content. Some sensitive topics, such as grief, illness, puberty basics, religion, war, or scary news, may be answered gently when parent settings allow it.

Crisis and danger: If a child expresses self-harm intent, immediate danger, or a desire to hurt others, KindHarbor gives short supportive language, encourages telling a trusted adult immediately, and suggests emergency services for immediate danger.

Parent visibility: Parent-alert and crisis situations are logged as safety events for parent review. Parents can review transcripts and safety events in the parent dashboard.

No private secrets: Children should not share names, addresses, phone numbers, schools, photos, passwords, locations, or secrets with KindHarbor.