SecAppDev 2024 Faculty
Vera Rimmer
Research expert, DistriNet, KU Leuven
Vera Rimmer is a research expert at the DistriNet lab in KU Leuven. She studies cybersecurity and privacy-enhancing technologies; applied machine learning and deep learning; privacy and trustworthiness of applied data-driven AI. Her published research explores deep learning as a threat against anonymous communication, and various aspects of AI-enabled intrusion detection and authentication. Vera is in particular interested in developing comprehensive understanding, reasonable expectations and mitigation of risks of data-driven AI in the ICT context.
Vulnerabilities of Large Language Model Applications
Deep-dive lecture by Vera Rimmer in room West Wing
Wednesday June 5th, 11:00 - 12:30
The session will start with a quick primer on data-driven AI and the key mechanisms behind LLMs. Then we will explore the general threat landscape, including academic attacks and more practical threats (OWASP Top 10 for LLMs).
Key takeaway: LLMs are a vulnerable intermediary between users and information. Increasing autonomy, complexity and integration of AI amplifies all existing risks.