Safety
Lamp is an AI
Lamp tells you it's an AI guide on its first screen and when voice starts, and it reminds you every three hours of continued use. If you ask, it will say so plainly. It never claims a life or memories of its own.
When someone mentions suicide, self-harm, abuse, or danger
- Typed messages are screened before any AI sees them. If one signals suicide, self-harm, abuse, or a threat of violence, Lamp doesn't send it to the AI. It shows a safety card instead.
- Spoken words are screened the same way as soon as they are transcribed, and the same card appears.
- The card leads with care ("I'm glad you told me. You matter, and this needs more than me."), asks whether you are safe right now, and gives one-tap links: 911; 988 (Suicide & Crisis Lifeline, call or text); text HOME to 741741 (Crisis Text Line); and for abuse, the National Domestic Violence Hotline, 1-800-799-7233 (or text START to 88788).
- After that, Lamp stays safety-first. It checks on your safety before anything else, and it doesn't quote Scripture at you, debate, minimize, ask for details, or give methods.
- Lamp never claims to be an emergency responder, a counselor, or God.
What we count
We count how many times the safety card is shown (not what was said), so we can report it where the law requires.