OpenAI released new tools to combat child sexual abuse material, including an image classifier that detects CSAM in both real and AI-generated images. The company developed the classifier with the National Center for Missing and Exploited Children and is providing access to select companies through its API at no cost.

In early testing with partners, the classifier flags 48% fewer images while finding the same number of abusive images, reducing reviewer workload. Images flagged by the tool are twice as likely to be abusive.

OpenAI is adding warning screens in ChatGPT that alert users attempting to access child sexual abuse material, direct them toward specialized support, and suspend accounts for 24 hours. Repeated violations result in permanent account removal. The company formed these messages and guidelines with organizations including Protect Children, the Lucy Faithfull Foundation, and Joel Scanlan at the University of Tasmania, who independently evaluated a trial of deterrent messages.

The company is publishing developer guidelines to help prevent, detect, and report CSAM content. In early testing, the guidelines helped developers adopt child safety tooling and register to report CSAM to NCMEC.

OpenAI is also providing law enforcement with AI-powered features including dedicated plugins, investigative tools, complex data analysis, training for investigators, and a field guide on using AI.

The move follows lawsuits against OpenAI and criticism from Common Sense Media, which labeled the company's technology as an "unacceptable risk" for teens.