Anthropic Updates Usage Policy to Combat AI-Generated CSAM and Child Safety Risks

Anthropic published an updated Usage Policy effective November 12, 2026, establishing stricter prohibitions on AI systems that compromise child safety. The policy explicitly forbids creating, distributing, promoting or facilitating child sexual abuse material including AI-generated CSAM, facilitating trafficking, sextortion or exploitation of minors, enabling grooming including content designed to impersonate a minor, and facilitating any form of child abuse including instructions to conceal it. Anthropic stated that when it detects CSAM, AI-generated CSAM, activity facilitating CSAM production or dissemination, or scenarios where a minor may be at imminent risk of sexual abuse or exploitation, it will report findings to relevant authorities, and has issued detailed Child Safety Guidance for Developers to help identify and mitigate child safety risks.

Platform policy changes that close down one generative pathway for abuse are necessary but not sufficient when offenders move from public prompts to private direct messages. Guardii's real-time detection architecture addresses the operational environment where AI-generated child sexual abuse material is most often weaponised: in one-to-one or small-group conversations on messaging platforms including Instagram, Snapchat, Discord and Roblox. The anti-CSAM module flags AI-generated and deepfake material at the point of transmission, surfacing the threat to a parent or authority before the child is exposed, and detecting escalation patterns that indicate sextortion, grooming or coercion regardless of whether the material was generated by a frontier model or sourced elsewhere, offering point-of-contact intervention where policy enforcement alone cannot reach.