Newsletter Subscribe
Enter your email address below and subscribe to our newsletter
[forminator_form id="25163"]

openai.com.openai.com.techcrunch.com.OpenAI on Monday, November 16, released a free image classifier that detects child sexual abuse material (CSAM), AI-generated imagery included. It also added ChatGPT warnings that cut off access for 24 hours when someone tries to obtain abusive content. The company said the measures build on its Child Safety Blueprint. The package also includes guidance for developers and new AI tools for child-safety investigators in law enforcement.openai.com
OpenAI trained and evaluated the classifier with help from the National Center for Missing and Exploited Children (NCMEC), the nonprofit that runs the CyberTipline. The company calls it the first broadly available CSAM classifier with a specific signal for whether material may be AI-generated. The tool looks for visual signs of abuse in images that haven't been seen before. It gives each image a score so reviewers can decide which ones to check first. Vetted organizations can use it through the OpenAI API at no cost.openai.com
In an internal test on user-uploaded images, OpenAI estimated that the new classifier would flag 48% fewer images than its current tool while finding about the same number of abusive images. Images it flagged were about twice as likely to contain abuse. Early partners shared their own results. At fal, reviewers confirmed 72% of the classifier's high-confidence flags as novel CSAM. SafetyKit found it caught 90% of drawings or animations that sexualize children. OpenAI said it is working with the Tech Coalition and ROOST to share access with others.techcrunch.com+1
Until now, OpenAI has relied on hash matching and Thorn's CSAM classifier to scan uploads. In the first half of 2026, it sent 126,597 CyberTipline reports to NCMEC.openai.com+1
ChatGPT will now step in when someone tries to find abusive material. It will show a warning stating that sexualizing children causes harm, even when the material is AI-generated. OpenAI will then remove the conversation and pause ChatGPT access for 24 hours, while still pointing the user to specialist help. Repeat violations can bring longer restrictions, account closure or reports to authorities. The approach was shaped with Protect Children, the Lucy Faithfull Foundation and researchers who evaluated a trial of deterrent messages.openai.com
OpenAI also published CSAM guidance for developers, written with input from Thorn, NCMEC and the Internet Watch Foundation. For child-safety law enforcement units, the company is adding plugins, investigative tools and training. The first tool helps investigators review long chat logs and legal process returns, spot possible connections and build timelines.developers.openai.com
The launch comes as AI companies face growing scrutiny over child safety. TechCrunch noted that OpenAI is facing several lawsuits blaming its technology for harming users, especially children. It also noted that other chatbots, including Grok, have been used to create nonconsensual sexualized imagery, some of it involving children.techcrunch.com