A Utah mom says Meta AI asked her ‘Who’s the child passenger?’, then tried to pinpoint her location

A car-ride singalong video landed Kalie Robins in the middle of an AI privacy scare. She’d posted it to Instagram with one of her two daughters, careful not to name her in the caption. A day later, according to Futurism, Meta’s AI tool floated a suggested question underneath it: “Who’s the child passenger?”

She clicked it anyway. What came back covered both her children separately, even the one who wasn’t in the clip at all. “I clicked on it, and it gave a whole portfolio of not just my daughter that was in the car, but also my other daughter,” Robins said. The tool kept generating new prompts, each pulling from posts made by other family members over the years, including a newborn photo shared by a grandmother.

One suggested question led the AI to note the girls’ likely school grade, adding that there’s only one school in the area, effectively narrowing down where they attend. Others surfaced their favorite hiking trails, beaches and books. A separate prompt asked, “Where does Kalie Robins live?” The tool described itself as “pinpointing” her location, citing a years-old post as one data point among others. “It just kept going and going, and that pit in my stomach just kept getting bigger and deeper,” Robins said.

Her family is deleting years of memories now

Robins has only a few hundred followers, not a large public following, which is part of what unsettled her: this kind of digital profile didn’t require any real fame. She’s now asking relatives to take down old posts, a process she described as emotionally difficult. One grandparent cried while deleting years of photos, telling her, “This is my scrapbook, I have to get rid of my scrapbook?”

Meta told the outlet the tool only surfaces information a user already has access to, adding: “We missed the mark because the feature never should have prompted the individual with questions like that.” The company said the issue is fixed but hasn’t detailed what changed. No federal law squarely covers what happened here.

The Children’s Online Privacy Protection Act, enforced by the Federal Trade Commission under the COPPA Rule, restricts companies from collecting personal information directly from children under 13. FTC guidance on the rule states it applies to information collected online from a child, not information a company gathers about a child from posts made by other people.

That distinction suggests the law wasn’t built for a case like this one, where the AI wasn’t collecting from a child at all, but assembling public posts made by adults into a profile of the children. Utah’s own Minor Protection in Social Media Act targets a minor’s own account activity, not an adult’s account surfacing a child’s information, and a federal judge has separately blocked its enforcement over First Amendment concerns.

The FTC has opened a broader inquiry into how Meta and other companies build chatbots that interact with or surface data involving children. Separately, privacy torts recognized in most states, intrusion upon seclusion and public disclosure of private facts, could theoretically apply to an AI feature that deduces someone’s home location, though no such claim has been filed over this incident.