OpenAI Launches Free CSAM Detection Tool and Toughens ChatGPT Penalties
Guardii Analysis
OpenAI announced on November 16, 2026 that it is releasing a new image classifier to detect child sexual abuse material including AI-generated content. The tool, developed with the National Center for Missing and Exploited Children, the Tech Coalition and the ROOST project, flagged 48 percent fewer images than current tools in early testing while detecting the same number of abusive images, and was approximately twice as likely to correctly identify abuse. The classifier will be made available free to vetted safety organizations. OpenAI is also introducing new warning screens in ChatGPT that explain the harm of sexualizing children, delete the flagged conversation, impose a 24-hour account suspension for attempts to access child sexual abuse material, and direct users to behavioral support resources, with repeated violations leading to permanent account removal.
Real-time detection at the point of contact addresses a gap that post-creation detection alone cannot close. Guardii's anti-CSAM module operates upstream, analyzing conversation patterns before imagery is requested, shared or created, intercepting grooming sequences that precede material exchange and sextortion attempts that use the threat of fabricated images as coercion. Guardii, backed by Startmate and a Meta Business Partner, identifies escalation and coercion in direct messages in real time across Instagram, Snapchat, Discord, Roblox and other platforms, alerting a parent, school or authority to a child in acute danger without constructing a readable feed of what was said, enabling intervention before an offender solicits or generates abusive material.
