Child safety
How we protect children while they use Momo
MiniGenius AI is designed as a parent-guided educational experience. We combine data minimisation, usage limits, curated content, and automated controls, while recognising that no artificial intelligence system can guarantee perfect results in every situation.
Who the service is for
The service is intended for families and must be activated by an adult parent or guardian. The main experience is designed for children aged 4 to 8; the 8-9 and 10+ ranges adjust difficulty and language but never expand the safety perimeter or permitted topics.
Momo is an AI-powered character
Momo is not a person, teacher, therapist, or emergency service. Some answers are generated through artificial intelligence models, which may make mistakes, misunderstand speech, or produce an unexpected answer. Children should understand that they are talking to a digital character.
Parents remain the reference point
The parent creates the account, gives consent, configures profiles and limits, and supervises use in a way appropriate to the child's age and maturity. Sensitive actions, billing, and privacy controls remain in the adult area.
Age-appropriate content
Momo uses short answers, simple words, concrete examples, and a non-judgmental tone. Age changes educational difficulty, not the rules concerning violence, sexual content, self-harm, personal data, dangerous activities, or other unsuitable material.
Child data is kept to a minimum
A profile uses a nickname, age range, language, avatar colour, and limits selected by the parent. We do not ask for a real name, school, city, address, phone number, photos, or free-text notes. Conversation audio and text are processed to provide an answer, but MiniGenius AI does not retain them as recordings or plain-text transcripts.
Multiple safety layers
Requests pass through preventive rules, automated moderation, Momo-specific instructions, and response checks. These systems seek to block or redirect unsuitable content, requests for personal information, and unsafe instructions. They reduce risk but do not remove the need for supervision.
When a question is delicate
For adult topics, violence, health, intense fear, pain, or signs of distress, Momo limits its answer and encourages the child to involve a trusted adult. If a critical signal is detected, we may record a safety event and alert the parent without including the child's words in the email. In a real emergency, contact an adult who is present or the emergency services immediately.
Games and stories with clear boundaries
Structured games and bedtime stories come from curated static catalogues. Bedtime stories are not generated live. Draw with Momo uses only paper, pencil, or drawing in the air; it does not suggest scissors, glue, or risky tools.
Controls for parents
Parents can manage profiles, age range, language, daily limits, subscription, and profile deletion. Only one child session may be active in a browser or device at a time. The dashboard shows duration, message count, aggregated topics, and any safety indicators, not a word-for-word conversation history.
Limits to understand
Speech recognition, moderation, and generative AI are not infallible. Momo may misunderstand a phrase, repeat itself, or provide inaccurate information. It must not be used for diagnoses, medical or psychological advice, emergencies, important decisions, or as a substitute for relationships with adults and peers.
Reporting and continuous improvement
Parents can report safety issues or inappropriate content by writing to hello@minigenius.ai with the information needed to identify the event, while avoiding unnecessary child data. We review controls, catalogues, and incidents to improve the service.
Last updated: 16 June 2026.