Lifecycle of a Phone Fraudster: Exposing Fraud Activity
summary
Analysis of 105 million call records finds about 1 fraudulent call per 2,900 (≈30,000 frauds) and $59M in losses, with attackers using ANI spoofing, VoIP gateways, prerecorded PBX hops, pitch shifting APIs and IVR probing. Acoustic fingerprinting of 147 features separates devices and networks and achieves 80% detection with <2% false positives, while graph/CDR modeling flags 46% reconnaissance cases 10–16 days before takeover, revealing multi-stage gangs and five-call takeover patterns.
AI-written summary · about
chapters (33)
- 0:00Introduction
- 0:38Overview
- 2:55Call Center Fraud
- 4:27IVR
- 5:40Call Center
- 5:57Reveal
- 7:26Mode of Operation
- 12:15Analysis
- 13:35Voice over IP
- 14:54Shortterm energy
- 15:50Packetization
- 16:20IRS Scam
- 17:58Electrical Grid Interference
- 18:46Acoustical Anomalies
- 20:52Distribution of Fraudsters
- 22:31Anomalies
- 23:51Separating Good and Bad Calls
- 24:44Creating Fraudster Profiles
- 25:34Plotting Fraudster Profiles
- 26:22Voice Distortion
- 28:10Red Pockets
- 29:18CDR Data
- 33:28Reputation Velocity Behavior
- 35:44Reputation Example
- 36:40Velocity Example
- 38:29Reconnaissance Example
- 40:18Overall Result
- 41:45Example
- 42:50Key takeaways
- 44:26Time gap between fraud stages
- 47:23False positive rate
- 50:29Changing vectors of attack
- 52:09How common is this technology
related talks
53mCall Me: Gathering Threat Intelligence on Telephony Scams to Detect Fraud
49mChipmunk or Pepe? Using Acoustical Analysis to Detect Voice-Channel Fraud
50mWinning the Battle Against Scam Calls – How We Mitigated Caller ID Spoofing
37mYour Phone is Not Your Phone: A Dive Into SMS PVA Fraud
09mModern fraud: A deep dive into Vishing
39mSynthetic Trust: Exploiting Biases at Scale
20mFraud fighting: leverage real-time network intelligence
20mSecurity in the Age of AI: Response and Strategy