OpenAI CSAM Image Classifier (omni-moderation-csam-latest)
OpenAI released an image classifier that detects potential CSAM, including AI-generated material, free to vetted organizations through the API, with new ChatGPT warnings and 24-hour pauses.
The classifier looks for visual signs of abuse and returns separate scores for potential CSAM and potential AI-generated CSAM, so reviewers can decide which images to look at first. OpenAI says it is the first broadly available CSAM classifier with a specific AI-generated signal, and it was trained and evaluated with help from NCMEC. Separately, ChatGPT now shows a warning to users seeking abusive material, removes the conversation and pauses the account for 24 hours.
- Date
- Monday 16 November 2026
- Lab
- OpenAI
- Kind
- product
- Access
- closed API
- Price
- No cost for approved organizations
Figures
| Measure | Value | Measured by |
|---|---|---|
| Images flagged versus prior approach | 48 percent fewer fewer false positives; flagged images were twice as likely to contain abusive material | company |
Access is limited to vetted organizations approved by OpenAI.
Sources
- openai.com/index/strengthening-online-child-safety
- developers.openai.com/api/docs/changelog
- theverge.com/policy/1034527/openai-csam-classifier
This record was partly confirmed: some claims could not be checked on 16 November 2026. How we check