Think like a fraudster: Reverse-engineering identity fraud

Fraudsters buy real driver's license images for as little as 25 cents, generate fake IDs with AI, and pay real people to pass selfie checks. Three people who fight them at Persona explain how each attack works and how to stop it.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
WHAT WE COVER

→ Three ways fraudsters get identities: data breaches, identity farms, and generative AI
→ Why mDLs, NFC, and authoritative databases are harder to fake than a physical ID
→ How fraud rings switch ID layouts and use virtual camera tricks
→ Why AI-generated IDs cluster tightly together while real IDs spread out
→ How portrait age vs. ID age exposes risky devices
→ How device risk score catches identity mule rings
→ How a deepfake injection pipeline works
→ How AI lets one fraudster act like an army
→ What the naked eye misses that a model catches, and the reverse

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

CHAPTERS

00:00 Welcome

01:34 Three ways fraudsters get identities

03:52 Catching synthetic identities

05:53 How fast fraud changes

07:53 Fraud controls and Signals Model Creator

10:16 Why AI-generated IDs cluster together

13:23 Portrait age vs. ID age

14:36 Identity mules and the $1,000 iPhone

16:28 How to measure if a fraud signal works

17:42 Inside Persona's red team

18:26 How deepfake injection works

19:31 How AI turns one fraudster into an army

22:17 Naked eye vs. machine

23:41 Where to reach the speakers

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

SPEAKERS

Ashley Fang, Fraud product analyst, Persona
Cathy Chang, Product manager, research and development, Persona
Jared Nagano, Identity red team, Persona
Hosted by Brandon Chen, product marketing manager, Persona

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

#IdentityFraud #Deepfakes #FraudPrevention #IdentityVerification #FraudDetection #KYC #TrustAndSafety #RedTeam