Companion AI Safety Protocol

AI identity & user awareness

LumiPet identifies its chat companions as AI in the chat experience. LumiPet’s AI Chat privacy disclosure explains that companions are artificial, not human, and may not be suitable for some minors.

AI-generated replies can be incomplete, inaccurate, or unsuitable for a particular situation. Use your own judgment.

If you are in crisis

LumiPet is not a crisis service, and its companions are not therapists, doctors, or emergency responders. If you are in immediate danger or thinking about harming yourself, please contact your local emergency services or a local crisis line right away, and reach out to a trusted person who can be with you. In the United States, you can call or text 988 (Suicide & Crisis Lifeline). If you type words that suggest you may harm yourself, the app shows a message with crisis resources before the conversation continues.

Safety boundaries

LumiPet’s companions are configured with instructions intended to keep conversations safe. They are designed to avoid encouraging self-harm, sexual content, content involving minors, serious violence, illegal activity, hateful content, manipulative dependency, and pressure to make purchases. They are instructed not to claim to be human or to be a therapist, doctor, emergency authority, lawyer, or financial adviser.

No automated system is perfect, and a companion may sometimes respond in ways that fall short of these boundaries. If that happens, please tell us.

Safety review

LumiPet may review reports you send us and related conversation records as needed to investigate a safety concern and improve the service, as described in the Privacy Policy. Retention is limited by the Privacy Policy and applicable law.

Report a safety concern

If LumiPet produced a dangerous or inappropriate response, or you want to report a companion safety issue, email us with the subject “Report an AI reply”: