How we use AI

Every Fablebee story is written by an AI model from the choices you make, and read aloud by a synthetic voice. No human writes or records your story.

The audio file itself also carries a marker inside it recording that it was generated by AI. It is there for software that checks whether audio is synthetic — it is not visible when you play the file, and anyone with a tag editor can remove it.

If you send a story to Yoto, Yoto converts the audio to its own format, and that conversion removes the marker. Instead, every story we send to Yoto carries the words “Created with Fablebee — AI-generated story” in the playlist’s own description, where anyone using your Yoto account can read them.

Each narrator voice is produced by a speech model from a short public-domain recording by a volunteer reader. No person records your story, and the narrator you hear is not a real person speaking.

Every story is checked automatically against safety rules written for young children once it is written and before it is turned into audio. If it does not pass, we do not send it, and your story credit is returned.

That check is done by an AI model, and it is careful but not perfect — it shares some of the same blind spots as the model that wrote the story. If a story ever reaches you that you are unhappy with, please tell us at help@fablebee.com and we will look at it ourselves.

The model is given the first name or nickname and age range of each child in the story, the name and species of any pet you added, the kind of story and the mood you picked, how long you asked for, and any details you typed. It is not given your email address, your name, or any payment information.

The safety check sends the finished story back to the same provider as a second request, so the child’s first name goes to them twice for each story. We keep a reference number for both.

Fablebee does not use your child’s name or data to advertise to them, to encourage them to keep using it, or to build a profile of them.