OpenAI releases a new image classifier and ChatGPT warnings to combat CSAM
OpenAI says its new classifier found the same number of abusive images while flagging 48% fewer images in early partner tests.
The classifier detects both real and AI-generated CSAM, with free API access for select companies. Developed with NCMEC, it scores previously unseen images to help reviewers assess them. ChatGPT will display deterrent warnings, direct users toward specialized support and impose 24-hour account suspensions for attempts to access material that sexualizes children; repeated offenses could lead to permanent removal. OpenAI also announced child-safety guidelines for developers and investigative tools and training for law enforcement.
TechCrunch AI's note
The classifier detects both real and AI-generated CSAM, with free API access for select companies. Developed with NCMEC, it scores previously unseen images to help reviewers assess them. ChatGPT will display deterrent warnings, direct users toward specialized support and impose 24-hour account suspensions for attempts to access material that sexualizes children; repeated offenses could lead to permanent removal. OpenAI also announced child-safety guidelines for developers and investigative tools and training for law enforcement.
TechCrunch AI's note
score 5