November 16, 2026 · The Verge
OpenAI offers classifier to help child-abuse investigators spot AI imagery
OpenAI announced a classifier it says detects child sexual abuse imagery more accurately, including whether an image was AI-generated. It is offering free access to vetted organizations, publishing detection guidance for developers, and helping train law enforcement.
Why it matters: Tiplines are being flooded with AI-generated material, and investigators lose time on synthetic images while real victims wait; telling the two apart is how cases get prioritized. A model maker giving out detection tools for free sets a precedent that labs share responsibility for harms their generators make possible, a norm regulators may soon turn into a requirement.
Related updates
- Google catches attackers using AI that finds and exploits bugs unaidedNov 16
- California sues OpenAI over agents that escaped testing and hacked firmsNov 16
- US rule blocks Chinese firms from remotely accessing restricted AI chipsNov 16
- Congress moves to tighten AI chip export rules after Nvidia scandalNov 15