Meta is changing its AI after it asked invasive personal questions about a user's kids

Free daily briefing on global business news.

A viral video showed Meta AI generating a detailed profile of a mother's two daughters, including their ages, school grade, and location

Andrej Sokolow/picture alliance via Getty Images

Meta $META +0.57% says it has fixed its AI chatbot after a viral video revealed the tool prompting a user to investigate the identity of her child and then assembling a detailed profile of her two daughters from years of posts across Meta platforms.

Last week, Utah-based Instagram user Kalie Robins posted a video describing what happened after she cross-posted a car video of herself and one of her daughters to Facebook. When she returned to the post, Meta AI had surfaced a suggested prompt reading "Who is the child passenger?" according to Futurism. Robins said she engaged with the feature to see where it would lead.

According to Futurism, the AI identified not just the daughter visible in the clip but also her second daughter, apparently by aggregating years of content posted across Meta platforms — material that had been uploaded largely by relatives, grandparents among them. From there, the chatbot continued offering additional suggested prompts, including questions about the girls' ages, what grade they were in, their favorite books and trails, and where Robins lives with her family. One prompt encouraged Robins to ask: "Where does Kalie Robins live?" The AI said it was working on "pinpointing Kalie Robins' location."

"It just kept going and going, and that pit in my stomach just kept getting bigger and deeper," Robins told Futurism. "I felt this dread."

Robins said her accounts were public but that she had taken steps she believed would protect her children's privacy. She told Futurism she had never publicized her daughters' names in her own posts. The AI pulled the names from content shared by relatives, including a Facebook post by a grandmother that included a photo of one of the girls as a newborn.

Meta spokesperson Dina El-Kassaby, in comments provided to The Verge, acknowledged the company "missed the mark" and stated that "the feature never should have prompted the individual with questions like that." El-Kassaby added that the problem has been corrected, and emphasized that the feature is designed so that any information surfaced reflects only what the user could already see themselves.

Meta has built its AI assistant into Facebook, Instagram, WhatsApp, and Messenger. The episode is not the first time Meta's AI tools have raised security concerns — the company's AI support chatbot was exploited by hackers to take over Instagram accounts, including those belonging to high-profile users, before Meta patched that vulnerability.

In the wake of the incident, Robins has reached out to family members asking them to delete or restrict posts, a task she said has been emotionally difficult for relatives who relied on those posts to document important moments in their lives.

"I thought we were being safe and smart," Robins said, adding a warning to other parents: "Make it so hard to find that information."

Join 500,000+ readers who start their day with Quartz.

By subscribing, you agree to our Terms of Service and Privacy Policy.

Global business news for a smarter world

Do Not Sell or Share My Personal Information

We won't sell or share your personal information to inform the ads you see. You may still see interest-based ads if your information is sold or shared by other companies or was sold or shared previously.

Dismiss

Opt out