OpenAI wants to stop deepfakes from overwhelming child abuse tiplines

The Verge
OpenAI announced a classifier to help child abuse investigators detect AI-generated CSAM and reduce false positives.

Summary

OpenAI announced a new classifier designed to help child sexual abuse investigators more efficiently identify harmful material, including AI-generated imagery of children. The tool reportedly reduces false positives by 48 percent while doubling the likelihood that flagged images contain abusive content, and it is the first broadly available CSAM classifier with a specific flag for AI-generated material. OpenAI is offering free access to vetted organizations, providing developer guidance, and training law enforcement. The announcement comes as AI-generated CSAM reports to the National Center for Missing and Exploited Children (NCMEC) CyberTipline have surged from 67,000 in 2024 to 400,000 in 2025, threatening to overwhelm investigators. OpenAI also added warnings and a 24-hour pause for users who attempt to generate CSAM with ChatGPT, and it reports all instances to NCMEC. North Carolina Attorney General Jeff Jackson called the collaborative prevention approach a novel experiment, while emphasizing the need for accountability.

(Source:The Verge)