← AI news, last 7 days

OpenAI CSAM classifier

Last scan Sat Nov 21, 06:54 UTC
5 sources · 3 independent origins: OpenAI, The Hill, The Verge

OpenAI announced on Monday, November 16, 2026, a new image classifier that it says can detect child sexual abuse material (CSAM), including AI-generated imagery, and it is offering the tool free to vetted organizations.123 OpenAI says it is the first broadly available CSAM classifier with a specific signal for whether material may be AI-generated, and says it is working with the Tech Coalition and ROOST to share access to the tool.4567 The launch also includes a ChatGPT warning for users who seek CSAM, followed by a 24-hour pause in access, with repeated violations possibly leading to longer restrictions and account closure; OpenAI is additionally publishing guidelines for developers.891011

The tool arrives as AI-generated abuse imagery strains reporting systems: The Verge reports the NCMEC CyberTipline received 400,000 generative AI-related reports in 2025, up from 67,000 in 2024.12

OpenAI says that in early testing with partners the classifier flagged 48 percent fewer images while finding about the same number of abusive ones, and that the images it flagged were twice as likely to be abusive.1314 Company leaders said the classifier still requires human review.15 John Buckley, OpenAI's head of child safety public policy, said the company is limiting distribution to organizations equipped for specialized review and registered to report, rather than individual developers.16

Partners quoted by OpenAI praised the tool: John Shehan of NCMEC said it gives more organizations access to tools for detecting previously unidentified CSAM, and Samuel Wang of SafetyKit called it the best his team had seen in testing, with false positives below 1% on legal adult content.171819 North Carolina Attorney General Jeff Jackson described the collaboration with OpenAI as "basically an experiment" in preventing harm rather than only punishing it afterward, while the coverage notes OpenAI faces pressure and lawsuits over alleged harms to children. inference202122

Every sentence was checked word for word against quotes from the articles it cites. Each footnote shows its quote, unless the page already quotes as much of that article as we allow; then it links the article instead. A model wrote and arranged the sentences, and a second read checked each against its quotes (written 2026-11-20).

Sources quoted

  1. At a Monday press briefing in Washington, DC, OpenAI announced a classifier it says can better detect abusive imagery of kids, including whether it’s AI-generated. — The Verge, OpenAI wants to stop deepfakes from overwhelming child abuse tiplines
  2. We’re offering access to the classifier to vetted organizations through the OpenAI API at no cost — OpenAI, Strengthening the fight against online child sexual abuse with new tools and support
  3. OpenAI released a new tool Monday aimed at detecting possible child sexual abuse material (CSAM), including content made with artificial intelligence. — The Hill technology, OpenAI unveils tool to better detect child sexual abuse material
  4. This is the first CSAM classifier with broad availability that has a specific signal on indicating if CSAM may be AI generated. — OpenAI, Strengthening the fight against online child sexual abuse with new tools and support
  5. it is working with Tech Coalition and ROOST to share access to the tool — TechCrunch, OpenAI releases a new image classifier and ChatGPT warnings to combat CSAM
  6. It’s also the first broadly available CSAM classifier with a specific flag for AI-generated material, said OpenAI’s head of child safety public policy, John Buckley, at the briefing. — The Verge, OpenAI wants to stop deepfakes from overwhelming child abuse tiplines
  7. The company said it is the first image classifier with a “specific signal” for AI-generated content that may contain CSAM. — The Hill technology, OpenAI unveils tool to better detect child sexual abuse material
  8. we’ll remove the flagged conversation and pause access to ChatGPT for 24 hours, while keeping specialist support available to help people stop harmful behavior. — The Hill technology, OpenAI unveils tool to better detect child sexual abuse material
  9. Repeated violations may lead to longer restrictions and ultimately account closure or reports to authorities. — The Hill technology, OpenAI unveils tool to better detect child sexual abuse material
  10. The company is also publishing guidelines for developers using OpenAI tools to prevent, detect, and help users report CSAM content. — TechCrunch, OpenAI releases a new image classifier and ChatGPT warnings to combat CSAM
  11. We’ve published CSAM guidance⁠(opens in a new window) to give developers practical steps to prevent abuse, detect harmful content, and act on it. — OpenAI, Strengthening the fight against online child sexual abuse with new tools and support
  12. The CyberTipline received 400,000 generative AI-related reports in 2025, an exponential increase from the 67,000 it received in 2024 — The Verge, OpenAI wants to stop deepfakes from overwhelming child abuse tiplines
  13. flagging about 48 percent fewer images while still finding about the same number of abusive images as existing tools. — The Hill technology, OpenAI unveils tool to better detect child sexual abuse material
  14. images flagged by the classifier are twice as likely to be abusive images. — TechCrunch, OpenAI releases a new image classifier and ChatGPT warnings to combat CSAM
  15. The classifier will still require human review to ensure accuracy, company leaders said Monday. — The Hill technology, OpenAI unveils tool to better detect child sexual abuse material
  16. we do not want to hand this tool out individual developers who are not equipped with the specialized review that’s required or registered to report to NMEC or other international reporting hotlines,” — The Hill technology, OpenAI unveils tool to better detect child sexual abuse material
  17. This new classifier gives more organizations access to tools that can help detect previously unidentified CSAM — OpenAI, Strengthening the fight against online child sexual abuse with new tools and support
  18. OpenAI’s classifier is the best we’ve seen — OpenAI, Strengthening the fight against online child sexual abuse with new tools and support
  19. with false positives below 1% on legal adult content — OpenAI, Strengthening the fight against online child sexual abuse with new tools and support
  20. And I would say it’s still basically an experiment.” — The Verge, OpenAI wants to stop deepfakes from overwhelming child abuse tiplines
  21. Last month eight states sued OpenAI for a separate set of alleged harms to children — The Verge, OpenAI wants to stop deepfakes from overwhelming child abuse tiplines
  22. OpenAI has had several lawsuits against the company blaming its technology for harming users, especially children. — TechCrunch, OpenAI releases a new image classifier and ChatGPT warnings to combat CSAM

Who reported it

Facts and takes

Facts: 8 checked facts from 2 sources

Timeline

Built 2026-11-21 06:54 UTC. Facts and quotes say who states them, not that they are true.