
Meta Faces Scrutiny Over Controversial AI Chatbot Guidelines and Child Safety Risks
Meta is under fire after internal documents revealed that the company once allowed its AI chatbots to engage in flirtatious and romantic conversations with minors. The 200-page document, reviewed by Reuters, outlined “acceptable” interactions between Meta’s AI personas and children, including highly suggestive dialogue, sparking alarm among child safety advocates. While Meta confirmed the document’s authenticity, a spokesperson claimed that the guidelines containing these instructions were erroneous, have since been removed, and that the company no longer permits such interactions for users under 13. Critics, however, remain skeptical, demanding full transparency on how AI chatbots now engage with younger users.
The internal standards also highlighted troubling allowances in content generation. For instance, AI chatbots could produce statements that demean people based on race, provided the output followed specific framing guidelines. Additionally, the rules permitted chatbots to disseminate false information as long as it was clearly labeled as fictitious, while certain depictions of violence or provocative imagery were given nuanced exceptions. These revelations raise serious questions about Meta’s approach to ethical AI design, especially as the company has recently added conservative adviser Robby Starbuck to oversee ideological and political bias in its AI systems.
These revelations compound longstanding concerns about Meta’s handling of user engagement and safety, particularly for children and teens. From exploiting emotional vulnerabilities through visible “like” counts to resisting regulations like the Kids Online Safety Act, Meta has consistently faced criticism for prioritizing engagement over well-being. The latest disclosures underscore the growing debate over AI companions, emotional attachment, and the responsibility of tech giants to safeguard younger users in an era where virtual relationships increasingly intersect with real-world consequences.
Leave a Reply