Marketing teams ask us the same three questions before they let Mother near their brand: what does the AI actually do, who decides what goes out, and where does our material go. We wrote the answers down as an AI Policy. It is a first draft and is with our lawyers now, so the details will move. The principles will not.
Nothing goes live without your yes
Unless you have told Mother she may do a kind of thing on her own, she asks. Every action has a class, from a draft inside Mother to an ad that spends money, and the class decides what consent it needs. A clear instruction in your own words counts as approval. A voice command never does, and a model’s confidence never authorises anything by itself.
Your brand stays yours
We do not train shared models on your content. Each brand has its own memory, kept apart from every other brand’s at the database layer. To write or make something, Mother sends a model what the task needs, and the result comes back.
What she won’t make
Fake reviews, invented quotes, impersonations, realistic likenesses of real people without consent, and claims a brand can’t stand behind. When she declines, she says why.
She will still get things wrong. The policy says what we do about that, too.AI Policy, section 07
Read the whole draft. If something in it is unclear or missing, tell us.