EU Commission proposes mandatory client-side scanning under updated Child Sexual Abuse Regulation as Parliament debates privacy safeguards

Members of the European Parliament debated the European Commission's proposed Child Sexual Abuse Regulation, which would mandate client-side scanning of private messages and cloud-stored content to detect child sexual abuse material, including AI-generated and grooming content. The draft regulation requires platforms to deploy detection technologies for known CSAM via hashing, for new or previously unknown CSAM using classifiers, and for grooming through conversation analysis. Civil liberties committees raised concerns about the privacy implications of bulk scanning, the reliability of automated classifiers, and the potential for false positives at scale. Industry representatives testified that encryption remains technically compatible with pattern-based detection but that hash-based scanning of end-to-end encrypted content would require breaking encryption or implementing backdoors, a requirement several governments and technology firms oppose.

Conversation-level detection offers a technical pathway that satisfies the Commission's child-protection mandate without the systemic privacy trade-off of content scanning. When detection operates on the pattern of an exchange rather than its plaintext, identifying sustained targeting, escalation toward a meeting or image request, or the coercive rhythm of sextortion, the system can surface a child at risk without constructing a searchable archive of what was said. Guardii's anti-grooming and anti-sextortion modules analyse the shape and seriousness of a conversation in real time, flagging hostile contact to a parent, school or platform trust-and-safety team and enabling rapid escalation to law enforcement where the threshold for emergency intervention is met, a design that detects the threat Parliament seeks to address while preserving the confidentiality encryption is intended to protect.