ITHome · Published 2026-09-13
Meta changes AI prompts after a family video sparks invasive questions
After sharing a video of herself singing in the car with her child, Kalie Robins found AI suggesting questions about her family.
Questions beneath an everyday post
ITHome reports that selecting a prompt led Meta AI to assemble family information from older posts. The Verge clarifies that the clip was cross-posted to Facebook. Meta acknowledged the inappropriate suggestions and says it fixed them.
Accessible fragments can become a profile
Meta says responses use content the person asking can already access. Aggregation matters too: separate everyday fragments can produce a fuller family picture when combined. Access permissions and the consequences of combining information are distinct issues.
The fix addresses suggested prompts
This response does not establish that every privacy concern is resolved, or that AI bypassed private-post permissions.
Sources and notes
sources中文
Sources and notes
Retained summary and context
The account comes from media reports. Meta’s response addresses suggested prompts, not every privacy concern.
Sources and evidence · 2 links
Each link has a purpose and scope. Link counts do not establish independent confirmation; direct support applies only to the statements specified below.
This page keeps the original link and a short account. Follow the links below for the publisher’s full explanation.
Source timeline
- · Found the report via ITHome.
Collection record
First discovered: ; added to the library:
Collection and source publication have distinct dates.