Mark RadarMARK RADAR
About
EN
Sign in

AI Impostors Force Banks to Rethink Identity Checks

2 reports · First detected 2026-09-17 · Last active 2026-09-17

Payment fraud is shifting from stolen credentials and suspicious transactions to identity attacks that begin before checkout. Criminals use AI-generated faces, voices, synthetic identities and human-like behavior to pass checks designed to separate customers from bots. The change matters because voice authentication, selfies and behavioral biometrics can now validate a convincing impostor. PYMNTS Intelligence and issuer processor Paymentology say financial institutions need continuous, risk-based trust scoring and scoped payment tokens instead of a single point-in-time identity check.

On Sept. 17, 2026, PYMNTS highlighted findings from its September Payments Innovation Tracker. Bankrate said 40% of adults encountered a financial scam during the previous year. UK Finance reported that scam attempts across more than 100 million accounts at nine financial institutions rose 62% in 2025, while phishing surged 140%. PYMNTS Intelligence said unauthorized-party fraud represented 71% of incidents and dollar losses, up from 48% a year earlier. UK Finance also logged 248,070 authorized push payment fraud cases, with losses rising 19% to £576.4 million.

All Coverage

2 original reports

The Backstory

The history behind this event
AI-Generated Identities Fuel New Fraud Threat as Digital Payments Turn to Programmable Trustfirst seen 2026-03-16 · 2 reports · similarity 0.80

Generative AI can now mimic voices, faces and even typing rhythms and user behavior, gradually eroding the reliability of voiceprints, selfies and behavioral biometrics used by banks. As AI agents also gain the ability to make payments for customers, risk controls are shifting from a one-time determination of “who you are” to continuous assessments of devices, transaction context and behavioral intent. The change has implications for the foundations of trust in digital wallets and the financial system.

On March 16, 2026, Paymentology Chief Technology Officer Tim Joslyn advocated replacing binary authentication with real-time trust scoring, allowing low-risk transactions to proceed while subjecting high-risk transactions to additional verification. He also proposed using rule-based tokenization to restrict AI agents, such as imposing a limit of just $5. The following day, FIS Vice President Christine Hurtubise called for tokenizing PANs and names and establishing dedicated authentication standards for agents.

Mark Radar|MARK RADAR

If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →

All times are in Taipei time (GMT+8)