Safety: what Kidikia does, and what it cannot guarantee
Kidikia’s safety relies on rules enforced by the server: child safety guidelines, refusal of inappropriate topics, server-side checks on quotas and payments, and data kept separate for each account. No AI is 100 % safe; we are open about that.
Content
- Refusal of sexual, violent, frightening, hateful, illegal, drug-related or self-harm content.
- Questions about the body, health or emotions: factual, age-appropriate answers, without a diagnosis.
- When distress is expressed: KIA encourages the child to talk to a trusted adult.
- KIA identifies itself as an AI and does not ask for personal secrets.
Accounts, data and payments
- Each member can access only their own conversations (database access rules).
- Stars, roles and payments can only be modified by the server.
- Visitors without an account are identified by a signed anonymous identifier, without an email address.
- Payments are processed through Stripe; Kidikia never sees card numbers.
- Purchased stars are credited only after signed confirmation from Stripe, and only once.
Limits
AI filters may let an error slip through or wrongly reject a legitimate question. Communications are encrypted in transit (HTTPS); the server and AI provider must read the conversation content to generate a response, so this is not end-to-end encryption.
This page is an automatic translation of the French version, which is the authoritative version. Written and reviewed by the Kidikia team · Last updated on October 1, 2026