Buy tickets

24–25 March 2027 · Munich

Financial crime · 2 September 2026 · 5 min read

Fighting AI-powered fraud with AI

Deepfakes, synthetic identities and automated social engineering are changing the threat model faster than most detection stacks were designed to handle.

Abstract data visualisation of real-time fraud detection signals across financial transactions

The European Banking Authority's 2026 Risk Assessment Report identifies AI-related cyber risk as a growing concern, while also highlighting AI's transformative potential for risk management, process automation and fraud detection. Both statements are true at once, and that tension defines the current fraud landscape.

Attackers now industrialise what used to be artisanal: voice cloning at scale, synthetic identities assembled from fragmented data, phishing personalised per recipient, and probing automation that finds the weakest control in a payment journey.

The defensive answer is rarely a single model. It is behavioural analytics across sessions, real-time decisioning at the transaction edge, identity signals that survive deepfaked media, and investigation tooling that lets financial-crime teams spend their time on judgement rather than data assembly.

The measure of success is not only the fraud stopped. It is the friction avoided for the overwhelming majority of legitimate customers.

More reading