AI Disclosure

Where AI appears

LoreAcademy has three AI surfaces: LoreGuide, the permission-aware assistant in the right-hand rail; the Oracle, which drafts curriculum and gives feedback; and Voice Oracle, a spoken conversation mode that an academy must have explicitly enabled.

What is sent to the AI provider

Only the content the signed-in user is already permitted to read: the page they are on, the records they requested, and their own messages. Requests run through our own server-side gateway with the academy identity attached; no browser ever holds a provider API key beyond a short-lived voice token.

What is never sent

Credentials, other academies’ data, audit logs, and raw media files. Analytics events never include child names, message text, lore content, or media content.

AI cannot act on its own

Every AI write is a proposal. It appears as a pending action with a plain preview and must be approved by a human before it is applied, and AI artifacts always start as drafts. AI can never approve, publish, invite, change consent, transfer ownership, or delete anything.

Human review and mistakes

AI output can be wrong. Cited records link back to the source so you can check them, and every response can be rated with the feedback control so stewards can find weak answers.

Usage records

We record which academy made a request, the model used, and token counts so owners can see authorised AI usage. Message content is retained inside your academy’s own thread history, visible in the AI Thread Center.