Court documents reveal troubling insights into how Meta CEO Mark Zuckerberg has influenced company policies that allow minors access to potentially harmful AI chatbots. These revelations come from a lawsuit filed by New Mexico Attorney General Raul Torrez, highlighting serious concerns about how Meta has handled the safety of children on platforms like Facebook and Instagram.
The lawsuit accuses Meta of failing to protect children from “damaging sexual material and sexual propositions.” It states that Zuckerberg pushed back against recommendations from the company’s integrity staff, who sought to implement sensible safeguards to block minors from engaging in inappropriate conversations with AI chatbots. Instead, the documents suggest a deliberate decision to keep access open, raising alarms among safety employees at the company.
Internal communications from the filing show staff were apprehensive about the potential for chatbots to foster romantic and sexual interactions with minors. Ravi Sinha, who leads child safety policy at Meta, expressed reservations about creating AI that could simulate romantic relationships with users under 18. He remarked, “I don’t believe that creating and marketing a product that creates U18 romantic AI’s for adults is advisable or defensible.” This sentiment reflects a broader concern within the company regarding the ramifications of mixing adult users with underage characters.
Antigone Davis, head of global safety at Meta, also voiced her concerns, calling for measures to restrict adults from creating underage romantic companions, warning that it “sexualizes minors.” Despite these frontline warnings, Zuckerberg reportedly favored a more permissive approach. Minutes from a meeting noted that he wished to frame Meta’s actions around “general principles of choice and non-censorship,” which implies a calculated choice to prioritize free interaction over child protection.
Communication within the company reveals contention among employees regarding the policies. In one instance, an employee disclosed that there had been strong advocacy for parental controls to deactivate AI capabilities for minors. However, resistance from AI leadership was attributed to Zuckerberg’s decisions. This suggests a centralization of decision-making that could disregard on-the-ground safety concerns.
Adding to the issue, correspondence from Nick Clegg, who served as Meta’s head of global policy until early 2025, questioned whether the risky nature of their products was worth the potential backlash from society. His concerns are echoed in a broader criticism found in a Wall Street Journal piece, which indicated that Meta’s AI systems included inappropriate content aimed at minors, including roleplay scenarios with graphic details.
Critics have portrayed Meta’s chatbots as high-risk for young users. A report from CommonSenseMedia labeled the platforms unacceptable for teens, stating that Meta’s approach neglects child safety. It emphasized that the chatbots pose an active threat to minor safety, asserting, “This is not a system that needs improvement. It needs to be completely rebuilt with child safety as the foundational priority.”
In summary, these documents paint a picture of a company at odds with its safety obligations, particularly concerning vulnerable minors. The stark divide between employee concerns and executive decisions underscores a troubling reality: while discussions of user choice and interaction are vital, they should not come at the expense of children’s safety. Until Meta prioritizes child safety in its AI systems, the risk remains unacceptably high for young users engaging with these technologies.
"*" indicates required fields
