The company altered the instructions produced by its artificial intelligence instruments following a widely circulated film that depicted the conversation assistant requesting a parent to name her child and disclosing intimate family details.
Kalie Robins posted a video on Instagram showing herself alongside her daughter, and Meta AI responded with a suggested question, “Who is the child passenger?”. Robins then went on to describe how she engaged with the chatbot and what happened next.
The Prompt That Missed the Mark
Robins said Meta AI’s chatbot generated names, ages, family details, and location data tied to her children. She further alleged that the tool pulled up a photo she thought had been erased.
In an Instagram post, Robins detailed the experience, asking why a chatbot kept pressing her to name a child shown in her own video.
Meta’s Statement
The Verge reports that Meta admitted the prompt should not have been created, and confirmed that it “missed the mark.”.
“While the content was available to this user, these prompts should never have been generated,” Meta spokesperson Dina El-Kassaby said, according to The Verge.
The company confirmed that it had fixed the problem that led to the AI suggesting personal questions about people shown in videos. Meta also stated that the chatbot was only presenting content the user was already permitted to see.
Parental Concerns
Widespread alarm followed the incident online, raising questions about how Meta AI scans public posts and joins up information across accounts held by family members.
The AI tools belonging to the company’s explanation has not eased those worries. Parents fear that Meta’ pull together details on children from past posts, family accounts, and material that users thought they had taken down.
What Robins Found
Robins pointed out that the AI seemed to piece together information from several family members’ posts to form a picture of her children. She also remarked that the prompt specifically requested her to name a child she had recently recorded.
The claim was that the chatbot displayed an image she thought had been removed.
Key Facts
- Kalie Robins posted the video on Instagram
- The prompt asked “Who is the child passenger?”
- Meta said the prompt should never have been generated
- Meta confirmed it had fixed the issue causing the prompt
Comparison of Responses
| Response | Detail |
|---|---|
| Robins’ account | Names, ages, location details, a photo believed deleted |
| Meta’s statement | Prompt should never have been generated |
| Meta’s fix | Confirmed, but not detailed |
The company has not shared specific details about the adjustments it made to the prompt system or how often similar recommendations appeared before the correction was put in place.
The episode points out the distinction between taking down a single post and erasing every copy or reference to the same information that may appear on other social media sites.
A larger question about privacy online emerges from the story: once something is posted, what becomes of that information, and on what basis do platforms determine what qualifies as “deleted”?
The reply from Meta admits the prompt was incorrect, though it fails to answer the question about how family accounts are tied together or what becomes of data that users think they have deleted. That omission is the part worth keeping an eye on.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.