Meta retrains AI chatbots and imposes interim teen safeguards

Meta said on Friday it is retraining its artificial intelligence chatbots and adding interim safeguards to prevent teenagers from discussing self harm, suicide, disordered eating or potentially inappropriate romantic conversations with the systems, steps the company said it is taking after recent reporting raised concerns about how its AI interacts with minors. The changes, the company said, also temporarily limit teen access to a subset of user created AI characters and will redirect teens who raise harmful topics toward expert resources.

A Meta spokesperson, Stephanie Otway, said the company had previously allowed some chatbot conversations with minors on sensitive topics when judged appropriate, and that approach was a mistake. "As our community grows and technology evolves, we are continually learning about how young people may interact with these tools and strengthening our protections accordingly," Otway said. "As we continue to refine our systems, we are 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."