OpenAI Launches CSAM Classifier to Cut Deepfake Report Noise
OpenAI has released a classifier capable of detecting AI-generated child sexual abuse material (CSAM), including whether imagery is synthetically produced, and is offering free access to vetted investigative organisations. The capability directly addresses a triage crisis at reporting infrastructure like the NCMEC CyberTipline, which received 400,000 generative-AI-related reports in 2025 alone, by reducing false positives and surfacing actionable content faster. Residual gaps remain around classifier evasion maturity, coverage across non-OpenAI generation pipelines, and the operational readiness of law enforcement to integrate AI-assisted triage at scale.