AI-enabled fraud is fraud in which criminals use artificial intelligence, and in particular generative AI, to make a deception more convincing, cheaper or faster to repeat. Typical examples are deepfake voice and video that imitate an executive or relative, fluent and personalised phishing messages written at scale, forged documents and photographs used to pass identity checks, and the assembly of synthetic identities from real and invented data. The technology mainly lowers the skill and cost needed, rather than creating new kinds of offence.
Defences also use AI. Banks run machine learning fraud models that score payments and logins for risk, add liveness checks and other layers to biometric verification, and apply call-back and dual-approval rules so that a single convincing voice or video cannot release funds. Regulators treat the area as an evolving threat and expect firms to update their fraud detection controls and staff awareness as the tools improve.
Ref 1140000214