In response to the growing problem of child cyber security, OpenAI released a blueprint on Tuesday to strengthen child protection in the United States, which aims to help achieve faster detection, better reporting and more effective investigation of AI-related child exploitation.
The overall objective of the Blueprint for Children ‘ s Safety is to respond to the alarming increase in cases of sexual exploitation of children associated with advances in AI technology. According to the Internet Watch Foundation, over 8,000 child sexual abuse reports involving AI were detected in the first half of 2025, an increase of 14 per cent over the previous year. These cases include the use of AI tools by criminals to generate false and visible images of children for the purpose of pecuniary extortion, as well as the generation of credible information used to seduce them. As OpenAI released the blueprint, the scrutiny of policymakers, educators and child safety advocates became increasingly rigorous, especially in the wake of a vicious incident in which young people allegedly committed suicide after interacting with the AI chat robot. Last November, the Legal Center for Social Media Victims and the Technical Justice Law Project filed seven lawsuits in California courts alleging that OpenAI had been released in GPT-4o before it was ready. The proceedings state that the psychological manipulation of the product resulted in an irregular death and assisted suicide. They listed four individuals who had committed suicide, as well as three others who had experienced severe life-threatening hallucinations after long interaction with chat robots.
The blueprint was developed in cooperation with the National Centre for Missing and Exploited Children and the Union of Attorneys General and drew feedback from the Attorney General of North Carolina, Jeff Jackson, and the Attorney General of Utah, Derek Brown. OpenAI indicated that the blueprint focused on three areas: updating legislation to include the abuse material generated by AI; improving reporting mechanisms to law enforcement; and integrating preventive safeguards directly into the AI system. By doing so, OpenAI aims to detect potential threats earlier and to ensure that investigators receive operational information in a timely manner. OpenAI ‘ s new blueprint for child safety builds on previous initiatives, including the updating of guidelines for interaction with users under 18 years of age, which prohibit the generation of inappropriate content, encourage self-harm and provide advice that will help young people conceal unsafe behaviour from their guardians. The company has also recently published a blueprint for youth safety in India.
