Child abusers exploit AI-generated deepfakes to blackmail victims into creating real abuse imagery, perpetuating a vicious cycle of sextortion that remains unchecked by global policing agreements.
AI systems are trained on vast datasets like Laion-5B, which may unwittingly contain illegal content like child sexual abuse material, highlighting the challenges of eliminating illicit training data sources.