# OpenAI offers free CSAM image classifier to approved organizations via Moderation API

> OpenAI's new omni-moderation-csam-latest model lets approved organizations screen images for child sexual abuse material at no cost.

- Status: official (From the company, authors or authority themselves, with a link to their own page.)
- Topic: Safety & Security
- Published: 2026-11-16T00:00:00+00:00
- Updated: 2026-11-16T17:09:08.832343+00:00
- Web page: https://ai-news.ikangai.com/story/openai-offers-free-csam-image-classifier-to-approved-organizations-via-moderatio/
- Seen in: OpenAI API changelog, Forward Future (Matthew Berman)
- Change: capability (New capability)
- Applies to: omni-moderation-csam-latest, v1/moderations

## What changed

OpenAI added omni-moderation-csam-latest under v1/moderations, a free image classifier for approved organizations that flags potential CSAM and potential AI-generated CSAM.

OpenAI has added a model that screens images for child sexual abuse material to its moderation endpoint. It is free for organizations the company approves. An entry in OpenAI's API changelog dated November 16 lists the model omni-moderation-csam-latest under v1/moderations and labels it a feature. The company says the OpenAI CSAM Image Classifier is meant to help approved organizations prioritize images for human review. It returns two separate signals: one for potential child sexual abuse material and one for potential AI-generated child sexual abuse material.

The documentation distinguishes this classifier from general moderation, which uses a different model. According to OpenAI's CSAM classifier documentation, each request takes exactly one still image, sent as an item in the input array with type set to image_url. The image_url.url field can be an HTTP or HTTPS address or a base64-encoded data URL. Images can be up to 20 MB (20,000,000 bytes). The classifier doesn't accept text, doesn't support inline moderation and doesn't work with the Batch API. Access requires an API key belonging to an approved organization, and organizations apply through a form. The documentation also tells them to follow their own approved input-handling process and not to submit known CSAM.

The response is a results array. Its first entry carries a flagged boolean and a categories object, with one boolean per category based on a configured decision threshold. Each category also gets a category_scores value from 0 to 1 with no units. The documentation warns against reading those scores as calibrated probabilities. It also warns against assuming 0.5 works as a universal threshold. Higher scores signal a stronger match within the same category and model snapshot. A flag does not establish that an image is CSAM. The classifier does not automatically report anything to authorities, so reporting remains the customer's responsibility under its own review and escalation processes. The docs' example sends an image of soap and returns flagged false, with likely-csam at 0.12 and likely-csam/gen-ai at 0.01.

The documentation draws a line between this model and omni-moderation-latest. That model can flag sexual content in images, but its sexual/minors category applies only to text. For image screening of this kind, the documentation points to the separate, access-restricted model rather than the general one.

## Why it matters

Trust and safety teams that screen user-uploaded images can add a dedicated classifier without paying for it, provided their organization is approved. The separate AI-generated signal targets a category that general moderation may treat differently.

## What you can do

Trust and safety leads handling user images should check whether their organization qualifies for approval to use the classifier.

## Original source

- [CSAM classifier | OpenAI API](https://developers.openai.com/api/docs/guides/csam-classifier)
- [Changelog | OpenAI API](https://developers.openai.com/api/docs/changelog)

## Further reporting

- [OpenAI Releases Child Abuse Image Detection Tools:](https://link.mail.beehiiv.com/ss/c/u001.KgoWwX0OhphnYTLatyBDCrUt5tX3UbW5gapJzkpAqiNWnbiKUtSGHWW-6yBpZlou_r7S871WkvB9O-4IOFCCQuQYNz9U11e8kEz--SJS9WRuerjXOalywl0Pno5fxibR0p4wHrwjTZULgjgcyoNY-iopkjhbVuOt1cw2Acf9OlvJmfbDoZmIcBeZcnKAAUv-LwPjM6oJ9GgZS5vNEY1rmaSomhXmfv9QQ6SO-O8Q0NN7pAyyLD0J5ny8PBbC7y7x0wnZUWFScmMBbkauGtLwQeDTunI8JmwaqBq28Xyq4-blSlutl6mq-r4G1h89SHtOw5Y2e2awEJoloXenjJwswTN8w1Ub2buK2eR7ZACIiao/4vt/caiZ5iTKQKuQ47nd67kTDg/h20/h001.xe66bOsr2I1ddTtTIxUYaR7M1nZk9krBnoRMVy11ORQ)

---

AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it.

Cite as: IKANGAI AI News, "OpenAI offers free CSAM image classifier to approved organizations via Moderation API", 2026-11-16, https://ai-news.ikangai.com/story/openai-offers-free-csam-image-classifier-to-approved-organizations-via-moderatio/
