Artificial intelligence (AI) is transforming criminal practice by industrialising deception, compressing attack cycles, and corroding evidentiary trust. This narrative review synthesises recent technical and criminological literature with institutional reporting to explain how generative models, predictive analytics, and automation enable convincing synthetic media, highly targeted social engineering, document forgery, identity synthesis, and adaptive evasion. Attention is given to the convergence with organised networks that use AI to coordinate logistics, mimic normal behaviour, and launder proceeds across platforms. Furthermore, a review of the grey literature was carried out to identify applied cases and to show how heterogeneous they are. Defensive efforts are advancing, yet detection remains brittle under laundering, increasing media realism, and adversarial adaptation. Regulatory and policy responses are surveyed across jurisdictions without claiming exhaustiveness; they appear fragmented and often lag operational innovation. The objective is pragmatic: to raise attacker costs and preserve information integrity while safeguarding fundamental rights and forensic reliability.
Bailo et al. (2026) studied this question.