Detecting Vulnerabilities in Generative AI

This project, a collaboration between 3Tenets Consulting Inc. and Dr. Wenjing Zhang from the University of Guelph, seeks to address emerging security and privacy vulnerabilities associated with the use of Large Language Models (LLMs) in enterprise environments. The initiative will focus on developing a prototype Privacy Leakage Assessment (PLA) Toolkit to evaluate and mitigate risks such as data extraction, membership inference, and prompt leakage attacks. Through systematic assessment, exploratory defense testing, and technical documentation, the project will provide 3Tenets with a preliminary framework to enhance its AI security offerings. This work supports the partner organization’s strategic goal of delivering advanced, privacy-aware cybersecurity solutions for clients adopting LLM-driven business applications.

Faculty Supervisor:

Wenjing Zhang

Student:

Partner:

3Tenets

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

University of Guelph

Program: