ChatGPT CSAM Warnings: 24-Hour Suspensions and New Classifier
OpenAI will pause personal ChatGPT for 24 hours after a child-safety warning, and it released a free image classifier for vetted organizations.

OpenAI will pause personal ChatGPT for 24 hours after a child-safety warning, and it released a free image classifier for vetted organizations.

On November 16, 2026, OpenAI published Strengthening the fight against online child sexual abuse with new tools and support ↗. Account rules are on the Help Center page, CSAM warnings and account suspensions in ChatGPT ↗. TechCrunch ↗ (Ivan Mehta, November 16) reported it. The Verge ↗ and TechRepublic ↗ (Aminu Abdullahi, November 17) also covered it. Figures below are from OpenAI's post unless the sentence names TechCrunch. The picture above is an original graphic created by aitoolranked. We have not used the classifier, and we have not seen one of these warnings.
OpenAI released an image classifier that can detect potential CSAM, including AI-generated material. It looks for visual signs of abuse so teams can find novel images, and it returns scores that help reviewers prioritize. OpenAI calls it "the first CSAM classifier with broad availability that has a specific signal on indicating if CSAM may be AI generated." It was built with assistance from the National Center for Missing and Exploited Children and is free to vetted organizations through the OpenAI API. The Tech Coalition and ROOST are helping share access through an intake form on the post. OpenAI names Thorn's Safer as another detection system it supports. The model card PDF ↗ is linked there, and this article does not summarize it.
The detection numbers are OpenAI's, from a comparison with its current protections on user-uploaded images. OpenAI estimated the new classifier would flag 48% fewer images while finding roughly the same number of abusive images, and that flagged images were about twice as likely to contain abuse.
Partner results in the same post are partner-reported, and OpenAI published them. They measure different things. OpenAI says fal reviewers confirmed 72% of high-confidence flags as novel CSAM. It says one in four flagged cases in invideo AI's testing led to a ban. It says SafetyKit detected 90% of cases involving drawings or animations that sexualize children.
OpenAI has not said how many ChatGPT accounts have been suspended. It has not published how the chat product detects an attempt, or a false-positive rate for that warning. The figures above cover the image classifier. The Help Center says the chat interventions roll out gradually worldwide. OpenAI has not given a completion date.
OpenAI says the warnings interrupt attempts to find abusive material, explain the consequences, and point people toward help. The warning states that sexualizing children causes harm, including when the material is AI-generated. OpenAI will remove the flagged conversation and pause access to ChatGPT for 24 hours, with specialist support still available. On repeat violations, the post says they "may lead to longer restrictions and ultimately account closure or reports to authorities."
OpenAI developed the warning with Protect Children, the Lucy Faithfull Foundation, and academics including Joel Scanlan of the University of Tasmania, who independently evaluated a trial of deterrent messages. The post also names Dr. Caoilte Ó Ciardha of the University of Kent among the advisors.
The Help Center applies this to Chat mode and Work mode, on web and mobile, for personal accounts. People using ChatGPT without signing in may also see warnings. It does not currently apply to Business, Enterprise, Edu, or Healthcare workspaces, and it is not available in voice mode, Codex, or the desktop apps. Rollout is gradual and worldwide. The Help Center says the interventions may appear differently by country. A warning stops the conversation. A signed-in chat is hidden from history and cannot be reopened. A suspension signs you out, and the email states why, how long, and which actions exist. After it ends you can sign in unless another restriction remains. A temporary suspension does not, by itself, mean permanent closure. OpenAI may close an account after a separate review. Appeal by the link in that email. An appeal does not restore access on its own.
OpenAI also published developer guidance with the Tech Coalition, Thorn, NCMEC, and the Internet Watch Foundation, and tools for child-safety investigators. The first tool reviews chat logs and legal process returns. With Anthropic and Google, OpenAI commissioned Tech Coalition guidance on disclosures of child exploitation in AI systems. TechCrunch, for its part, says OpenAI faces several lawsuits blaming its technology for harming users, especially children, that Common Sense Media labeled OpenAI's tech "unacceptable risk" for teens, and that other chatbots, specifically Grok, have been used to create nonconsensual sexualized imagery, including of children.
"Unfortunately, folks at frontier labs deal with real safety issues. This is one of them. The safety issue isn't all theater." — Ben Badejo ↗ (@BenjaminBadejo)
"Instead of wading through a pile of false positives, our trust and safety team gets a smaller, sharper set of cases to review, and a lot more of them turn out to need action." — an invideo AI statement published by OpenAI
"Our Stop It Now helpline already regularly hears from people who say an AI chatbot sent them to us." — Frances Frost of the Lucy Faithfull Foundation, in a statement OpenAI published
Reddit was not reachable for this story.
Personal accounts get the lockout. Workspaces, voice, Codex, and the desktop apps do not. A flagged chat leaves history and ChatGPT pauses for 24 hours. The email holds the appeal link, and an appeal does not restore access by itself. ChatGPT for Healthcare is outside this pause. Image hosts can request the free classifier. Thorn's Safer is the other system OpenAI names. Those scores are OpenAI's and its partners', and they cover the classifier, not the chat warning.
Watch appeals, coverage, and other labs' numbers. TechCrunch's lawsuits and the Common Sense Media label are the backdrop. California's rogue-agent suit is a separate sandbox case. ChatGPT vs Claude vs Gemini and the ChatGPT alternatives guide are the assistant pages on this site.
Verdict: a narrow pause on personal ChatGPT, and a free classifier scored by OpenAI and its partners.