Fraudsters live and die today by executing on what some may call — prank calls. Only the punchline hits businesses in their pockets, leaving law enforcement and companies to ask, “How do we know stop them?” They are robocalls, voice phishers and caller ID spoofers using cybercrime techniques to launch scam campaigns through the telephony channel that many people have long trusted.
I am excited for Marzuoli to discuss her latest research findings on the most menacing trends of the telephony channel and describe the calling patterns she tracked via a telephony honeypot. She will share with you her original thesis and how she used Pindrop’s honeypot to gather and analyze accurate and timely information on unwanted phone calls across the United States. By determining how these bad call sources can be quickly and accurately identified using features extracted from honeypot call audio, Pindrop Labs stands to aid law enforcement and businesses across the globe that are combatting rising telephony fraud.
Using machine learning and semantic information collected from honeypot call audio, Marzuoli and her team collected over 500,000 calls over five months from 90,000+ unique source phone numbers. Leveraging this data, Pindrop Labs developed a method to “fingerprint” high-risk call sources, attempting to hide behind phone numbers, and detect them in the first few seconds of a call.
In total, Marzuoli’s research included 1,072,840 calls placed to a honeypot by 209,755 sources and to 57,818 destinations. Out of these calls, she sampled over 100,000 for recording and analysis. The eye opening results were shared with Federal Communications Commission and the Federal Trade Commision. We look forward sharing to these results and what can be done to protect consumers and stop robocall scams.
I will be in the audience as she takes the stage and we will both be available after the session. I hope you will join us at Black Hat on August 4, 2016 at 5 pm PST in the South Seas CDF room. More details on her session are available here.