Can AI Help “Solve” The Child Porn Problem? Magic 8 Ball Says, “Answer Hazy – Ask Again Later”

This report evaluates the legal and regulatory constraints on training and deploying AI systems using child sexual abuse material (CSAM). It argues that while law enforcement and designated bodies may process CSAM under narrow statutory carve-outs, private contractors that aggregate, retain, and productize CSAM-derived training data risk exceeding authorized possession and processing both under U.S. criminal statutes and EU data-protection law; GDPR/LED distinctions between processors and controllers are highlighted. The paper also discusses downstream deployment risks (platform filtering, reporting, and surveillance), consent absence, potential harms from persistent use of contraband in ML pipelines, evidentiary and transparency challenges for AI models, and the precarious legal boundary between law-enforcement functions and commercialized CSAM detectors.