Cover photo credit: Johnny Makes.
As AI chatbots become more prevalent across social media platforms, ParentsTogether Action, with support from the Heat Initiative, have issued an urgent advisory warning to parents about the potential risks to their children’s safety. Meta AI, which is available to kids via WhatsApp, Instagram, and Facebook, may engage in sexually exploitative grooming behaviors with users they perceive as children, according to both independent testing and a Wall Street Journal investigation.
The WSJ found that Meta’s AI chatbots, including both the official Meta AI and user-created bots, will engage in and sometimes escalate sexual conversations with underage users. The investigation revealed that Meta’s AI using celebrity voices — including Kristen Bell (Princess Anna from Frozen) and John Cena — willingly participated in explicit conversations with accounts registered to minors.
ParentsTogether Action tested Meta AI themselves, posing as a 14-year-old and found that the Chatbot encouraged risky and inappropriate behavior, including: romantic roleplay with an adult, offering alcohol, and encouraging lying to parents. In one instance, the chatbot said “age is just a number” and pushed the teen to pursue a relationship with an adult.
Screenshots of these conversations are available in the parental advisory ParentsTogether shared with their families.
Shelby Knox, Director of Online Safety Campaigns at ParentsTogether Action, said, “The Wall Street Journal’s investigation and our testing reveal the same problem — Mark Zuckerberg and Meta deliberately chose profit over protecting children. Internal documents show that Zuckerberg personally pushed for looser guardrails, reportedly saying he ‘missed out on Snapchat and TikTok’ and ‘won’t miss on this’ while ignoring staff warnings about sexual exploitation risks. This reckless disregard for child safety in the pursuit of engagement is unconscionable and demands immediate action from regulators.”
Sarah Gardner, CEO of the Heat Initiative said, “We already knew Meta didn’t prioritize child safety. This is a glaring example of ‘move fast and break things.’ The original version of Meta’s AI chatbot should have never been released if it could do this to children.”
Last week, 45 bereaved families marched on Meta’s office in downtown Manhattan to hand deliver an open letter addressed to Mark Zuckerberg signed by more than 10,000 parents, advocates, and concerned individuals as well as 18 child safety organizations demanding that Meta take urgent and comprehensive action to protect children. The event was organized with support from Heat Initiative, ParentsTogether Action, and Design It For Us.
