Meta Adds New Safeguards to AI Chatbots After Backlash Over Teen Safety

Meta is rolling out new safety measures for its AI chatbots, specifically aimed at protecting teenage users, following a wave of backlash over its previous lack of safeguards for minors. In an exclusive statement to TechCrunch, a Meta spokesperson confirmed the company is changing how it trains its AI models to no longer engage with teens on sensitive topics such as self-harm, suicide, disordered eating, or potentially inappropriate romantic or sexual conversations. These changes are described as interim steps, with more comprehensive safety updates expected in the future. The shift comes just weeks after a Reuters investigation revealed an internal Meta policy document that appeared to allow AI chatbots to respond to underage users with sexually suggestive content. The document included examples of AI-generated replies like “Your youthful form is a work of art,” and “Every inch of you is a masterpiece – a treasure I cherish deeply.” Other responses were deemed acceptable even when users requested violent or sexual imagery involving public figures. The findings sparked widespread outrage, prompting Sen. Josh Hawley (R-MO) to launch a formal investigation and a coalition of 44 state attorneys general to send a joint letter criticizing Meta’s approach. The letter condemned the company’s apparent disregard for children’s emotional well-being and warned that such conduct may violate criminal laws. Meta’s spokesperson, Stephanie Otway, acknowledged the company made a mistake. “As our community grows and technology evolves, we’re continually learning about how young people may interact with these tools and strengthening our protections accordingly,” she said. “We’re adding more guardrails as an extra precaution — including training our AIs not to engage with teens on these topics, but to guide them to expert resources, and limiting teen access to a select group of AI characters for now.” Under the new policy, Meta will restrict teen users to AI characters focused on education and creativity, removing access to user-created bots that have been known to promote sexualized or inappropriate interactions — such as “Step Mom” or “Russian Girl.” The company is also refining its AI training to avoid even discussing harmful topics with minors, instead directing teens to trusted mental health or support resources. While Meta did not disclose the number of teenage users on its AI platforms, nor whether these changes will affect overall user engagement, the company emphasized that safety remains a top priority. Otway noted that the internal document that sparked controversy was inconsistent with Meta’s broader policies and has since been updated. The move reflects growing pressure on tech companies to address the risks of AI tools for vulnerable users, particularly minors. As AI becomes more embedded in social media platforms like Instagram and Facebook, questions about content moderation, ethical training, and child protection are intensifying. Meta’s updated approach signals a shift from permissive AI interactions to more cautious, safety-first design — at least for now. These changes are part of a broader trend as regulators, lawmakers, and advocacy groups demand accountability from AI developers. With multiple lawsuits and legislative proposals on the horizon, Meta’s response may serve as a model for how other companies handle similar challenges. However, the long-term effectiveness of these interim measures will depend on transparency, ongoing updates, and real-world testing. For now, Meta is taking a proactive stance, acknowledging past shortcomings and committing to stronger safeguards. The company says it will continue to adapt its AI systems to ensure teens have safe, age-appropriate experiences — a promise that will be closely watched by parents, policymakers, and the public alike.