SmishSmashing - A System Software Utilizing LLMs (GPT-4o) to Detect and Respond to Text Message (SMS) Phishing Scams
Published:
“FREE MINECRAFT DOWNLOAD”. “Your Car’s warranty has expired.” “Congratulations! You have just won a free iPhone!”. All of these are scams that Internet users—especially those less tech-literate—often fall for.
Adhering to the purpose of SEFCOM (The Laboratory of Security Engineering for Future Computing at ASU), SmishSmashing is intended to help make technology safer and protecting those who may be prone to cyberattacks—particularly smishing.
- Smishing is a type of cybercrime that fraudulently contacts victims via SMS (text messages) to reveal personal information, like passwords or credit card numbers.
SmishSmashing utilizes the most advanced large language models (LLMs), such as GPT-4o, to detect and respond to common text/SMS scams (”smishing”) with expressions, dialects, and behaviors that mimic a real human.
Curious about the development process? Check out the embedded Notion wiki attached below (or view the original website in a separate window)
View original website in a separate window
Summary Poster
Presented at the 2024 SEFCOM High School Internship cumulative poster session.