OpenAI Launches New Classifier to Detect AI-Generated Child Abuse Material
OpenAI has announced a new classifier designed to help investigators distinguish between real and AI-generated child sexual abuse material (CSAM). The tool aims to alleviate the massive influx of generative AI reports currently overwhelming systems like the National Center for Missing and Exploited Children (NCMEC) CyberTipline, which saw reports jump from 67,000 in 2024 to 400,000 in 2025. The new technology reportedly reduces false positives by 48% while doubling the likelihood that flagged images contain actual abusive material. OpenAI is providing free access to vetted organizations and offering guidance to developers to help mitigate the risks posed by open-source models and 'nudifying' apps. This move comes amid increasing regulatory pressure and legal challenges regarding AI safety and child protection.