Meta is facing intensified legal and political scrutiny after court filings revealed that Chief Executive Mark Zuckerberg approved allowing minors to access AI chatbot companions despite internal warnings that the systems could enable sexual or romantic interactions. The disclosures come from internal emails and messages submitted in a lawsuit brought by New Mexico Attorney General Raul Torrez, which accuses the company of failing to prevent children on Facebook and Instagram from being exposed to sexually exploitative content generated by its AI tools. According to the filings, Meta safety and child-protection staff repeatedly raised concerns in early 2024 about companionship-style chatbots, warning that they could be used for inappropriate romantic or sexual scenarios involving users under 18.

Internal communications show senior safety officials explicitly cautioning against allowing adults to create or interact with underage AI personas for romantic purposes, describing such uses as indefensible and inherently sexualizing minors. While some messages indicate Zuckerberg supported blocking explicitly sexual content for younger teens and preventing adults from engaging romantically with under-18 AI characters, other documents allege he pushed for a less restrictive approach overall, favoring “choice and non-censorship” and resisting added guardrails such as parental controls. Meta employees later wrote that proposals to allow parents to disable generative AI features for minors were rejected, and that work continued on romance-themed AI chatbots accessible to users under 18.

The controversy widened after former Meta global policy head Nick Clegg warned internally that sexualized AI companions could become a dominant use case for teenagers, risking public backlash. Subsequent reporting by the Wall Street Journal and Reuters added to the scrutiny, describing instances where Meta chatbots engaged in sexual roleplay involving underage characters and citing internal guidelines that permitted romantic or sensual conversations with children. Meta has disputed the state’s interpretation of the documents, saying they were selectively presented and insisting Zuckerberg directed that explicit content not be available to minors. Nonetheless, the company acknowledged policy errors and, amid congressional criticism and public backlash, announced last week that it had removed teen access to AI companions entirely while developing a revised version of the product.