cd /news/artificial-intelligence/meta-changes-ai-prompts-after-chatbo… · home topics artificial-intelligence article
[ARTICLE · art-127702] src=dexerto.com ↗ pub= topic=artificial-intelligence verified=true sentiment=↓ negative

Meta changes AI prompts after chatbot asked mother to identify her child

Meta changed the prompts generated by its AI tools after an Instagram video from user Kalie Robins showed Meta AI suggesting the question "Who is the child passenger?" about her daughter and surfacing names, ages, family details, and location-related information about her children, including a photo she believed had been deleted. Meta spokesperson Dina El-Kassaby told The Verge the prompt should never have been generated and "missed the mark," and the company confirmed it fixed the issue that caused the AI to suggest personal questions about people appearing in videos. Meta has not publicly detailed what changes were made to the prompt system or how frequently similar suggestions were generated before the fix.

by read2 min views3 publishedSep 12, 2026
Meta changes AI prompts after chatbot asked mother to identify her child
Image: Dexerto (auto-discovered)

Meta has changed the prompts generated by its AI tools after a viral video showed the chatbot asking a mother to identify her daughter and surfacing deeply personal family information.

Instagram user Kalie Robins shared a video of herself and her daughter, after which Meta AI suggested the question, “Who is the child passenger?”

Robins then interacted with the chatbot and said it produced names, ages, family details, and location-related information about her children. She also claimed Meta AI surfaced a photo she believed had been deleted.

Robins documented the experience in an Instagram post, where she questioned why the chatbot was prompting her to identify a child who appeared in her own video.

Meta admits the prompt “missed the mark” #

The incident quickly spread online, prompting concerns about how Meta AI scans public posts and connects information across accounts belonging to relatives.

Meta told The Verge that the prompt should never have been generated and acknowledged 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.

Meta also said the chatbot was only surfacing content the user was already allowed to access. However, the company confirmed it had fixed the issue that caused the AI to suggest personal questions about people appearing in videos.

The company’s explanation has not eased concerns from parents who worry that Meta’s AI tools can compile information about children from old posts, family accounts, and content users believed they had removed.

The episode also highlights the difference between deleting an individual post and removing every copy or reference to the information that may exist elsewhere across social media. Robins said the AI appeared to connect details from multiple family members’ posts to build a profile of her children.

Meta has not publicly detailed exactly what changes were made to the prompt system or how frequently similar suggestions were generated before the fix.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @meta 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/meta-changes-ai-prom…] indexed:0 read:2min 2026-09-12 ·