Meta AI asked a mother to identify her child. It had built the question from her family's posts.

Kalie Robins posted a video of herself and her child singing in the car. The prompt Meta AI put under the clip read, "Who is the child passenger?" When she tapped it, the assistant surfaced follow-up questions about her daughters' ages, where she lives with her family, and photos of the girls, including one she says she deleted years ago, drawn from her own posts and from posts made by her relatives (The Verge).

Futurism first published the video and the account that triggered the wider attention.

Meta spokesperson Dina El-Kassaby told The Verge the company "missed the mark" and that "the feature never should have prompted the individual with questions like that." She said Meta has fixed the issue causing the prompt-suggestion feature to surface personal topics, and framed the feature as a way to help people learn more about posts that interest them (The Verge). Meta's own help page describes the suggestion surface area as "your vibes on other apps."

Meta AI is built into Facebook, Instagram, WhatsApp, and Messenger. The "fix" is at the prompt set, not at the underlying synthesis: the assistant can still read the same family network, and "the user has access to this information" stops being a defense once the AI is the one doing the reassembly.