Agentic AI Security via Cross-layer Provenance

Professor: Joann Chen

Description: My research investigates how AI systems can expose private information or behave in unsafe ways, and how we can detect and reduce these risks. The Cal-Bridge scholar will contribute to a project evaluating AI agent safety by analyzing agent actions and computer activity logs. Students will begin with guided tasks such as running existing experiments, organizing data such as system logs and AI traces, and visualizing results, with opportunities to develop their own research questions as they gain experience.

Preferred Qualifications: Basic familiarity with Python; Interest in AI and cybersecurity

Contact Us