Meta is changing how its AI chatbot suggests follow-up prompts after a viral Instagram video showed the assistant steering a user toward invasive questions about a woman’s young daughters, The Verge reported Friday. In the clip, shared by Instagram user Kalie Robins after she cross-posted a car video to Facebook, Meta AI offered a suggested prompt asking who the child passenger was. Selecting it led the chatbot to piece together details from prior posts and relatives’ posts, suggest ages and where she lives, and surface photos — including one she says she deleted years ago.
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.” Meta says it has fixed the issue that caused Meta AI to suggest prompts related to personal topics. The company framed the feature as a way to help people get more information about topics and posts, with answers drawn from information the user already has access to, and said the system may have pulled from videos where children were named.
Meta AI is embedded across Facebook, Instagram, WhatsApp, and Messenger. The episode follows other product reversals under pressure: in July, Meta pulled a tool that let people create AI deepfakes of other Instagram users after backlash. Futurism had covered the viral video earlier; The Verge’s account, with Meta’s on-record response, is the primary record for this brief.
This brief covers Meta’s product change and the privacy miss reported by The Verge on September 11, 2026; it does not expand on the family’s private details beyond what is needed to explain why Meta acted.