OpenAI released a new AI image classifier built with NCMEC to detect CSAM, including AI-generated imagery, offering free API access to select companies. The classifier reduced reviewer workload by 48% while doubling the proportion of flagged images that are abusive. Additionally, ChatGPT now shows warning screens, directs users toward support resources, and imposes 24-hour temporary suspensions for attempts to create or access content sexualizing children.
Background
Child sexual abuse material (CSAM) detection remains a critical challenge as generative AI makes it easier to produce illicit synthetic imagery. NCMEC requires companies to report known CSAM to law enforcement, and OpenAI's classifier extends these efforts to previously unseen and AI-generated content.
- Source
- TechCrunch
- Published
- Nov 17, 2026 at 01:00 AM
- Score
- 7.0 / 10