Vadrevu Wins NSF CAREER Award to Turn Scammers' Games Against Them

The interaction seems routine: a friendly phone or video call or text or chat message from your bank, your computer "support team," or a long-lost match from a dating app. What follows is a slow, patient exchange engineered to separate you from your money. These scams are called interactive social engineering attacks. Instead of breaking into computer networks, scammers trick people into giving away confidential financial information like passwords or credit card numbers. The attacks run in real time, in two-way exchanges, and drain billions of dollars from victims every year. They’re maddeningly hard to track, study, and prevent because they unfold in private over phone or computer, one target at a time.

LSU Engineering's Phani Vadrevu is going after these social scammers. Vadrevu, an assistant professor in the Division of Computer Science and Engineering, has won a $630,000 National Science Foundation CAREER Award – among the most prestigious grants the agency gives early-career scientists – to build the first systematic defense against interactive social engineering attacks.

He will tap a global community of "scambaiters" – volunteers who deliberately bait scammers and waste their time. Vadrevu's team will mine the tools and tactics this community has developed to keep scammers talking, then pair those insights with controlled, ethically-safeguarded systems that engage scam operators directly.

The payoff is data that almost no one has – real conversational transcripts of scams in progress. From that data, Vadrevu’s research team will train computerized detection programs on three kinds of evidence: the real-time transcripts, the behavioral fingerprints of how victims use their computers, and the acoustic signatures of a scammer voices. The goal is to better understand how scammers persuade victims to share personal information – insights that will help detect and disrupt these attacks.

The project will also include public education. Vadrevu's team will develop an AI-based tool to simulate realistic scam experiences, which will be used in educational workshops for Louisiana residents. The goal is to increase awareness of these social attacks and how to avoid falling victim to them. Project leaders will partner with law enforcement agencies at home and abroad in these efforts.

Vadrevu is a member of the Division of Computer Science and Engineering’s growing bench of researchers focused on cybersecurity and holds a joint appointment in LSU’s Center for Computation and Technology. He earned his Ph.D. at the University of Georgia and came to LSU in 2023 from the University of New Orleans. He has presented at top security conferences, such as the USENIX Security Symposium and IEEE Symposium on Security & Privacy. He has received two previous rounds of National Science Foundation funding to study social engineering attacks.

“Antivirus software and firewalls can’t protect you from social scams,” said Ibrahim Baggili, the division chair. “The information and education that will result from Phani’s work, however, will help provide the protection the public needs.