KEY POINTS
- OpenAI announced a classifier that detects child sexual abuse imagery and flags whether it is AI-generated.
- The company says it produces 48 percent fewer flagged images, and those flagged were twice as likely to be abusive.
- NCMEC's CyberTipline received 400,000 generative AI-related reports in 2025, up from 67,000 in 2024.
OpenAI has announced a classifier meant to help investigators find child sexual abuse material more quickly, including content generated by AI. The company presented the tool at a Monday press briefing in Washington, DC, according to The Verge.
Alongside the classifier, OpenAI said it will give vetted organizations free access, publish more guidance for developers on detecting this kind of content, and help train law enforcement on how AI can support investigations.
The announcement comes as reports of AI-related abuse imagery grow quickly. The Verge reported that the National Center for Missing and Exploited Children's CyberTipline received 400,000 generative AI-related reports in 2025, compared with 67,000 in 2024. Researchers have warned the volume could become unmanageable for the center, and that a flood of fake images can make it harder to spot when a real child is being harmed.
OpenAI says the new tool reduces the burden on human reviewers. It claims the classifier flags 48 percent fewer images than before, while the images it does flag are twice as likely to contain abusive material. The company says it is built to catch AI-generated content from many sources, including open-source models and so-called nudifying apps.
John Buckley, OpenAI's head of child safety public policy, said it is the first broadly available classifier of this kind with a specific flag for AI-generated material. He argued that telling apart a fictional child from a real one is important for triage in law enforcement reporting. He also said laws should recognize the harm of synthetic material, which he said can erode social norms and is often found alongside imagery of real children.
OpenAI is also changing how its own products respond. ChatGPT's image generator is designed to refuse requests for such content, and users who try will now see a warning intended to encourage them to seek help. A first attempt triggers a 24-hour pause on access and a notice that the account could be permanently closed, along with support resources. OpenAI says it reports all instances of child sexual abuse material to NCMEC.
North Carolina Attorney General Jeff Jackson, who cochairs the AG Alliance's AI Task Force, called the growth in AI-generated reports "explosive" and warned that continued growth could mean authorities fail in their mission. He described the partnership with OpenAI as a way to prevent harm rather than only punish it afterward, but said it is "still basically an experiment."
The Verge noted that pressure on AI companies is increasing. The Take It Down Act requires platforms to remove nonconsensual intimate images, real or AI-generated, within 48 hours, and 48 states have laws on sexually explicit deepfakes. Last month, eight states also sued OpenAI over separate alleged harms to children.