Completed research · 2026
AdvPIIBench
An open benchmark for testing PII detection when sensitive values and their context are deliberately transformed.
AI privacy
Keep sensitive information protected.
How can AI systems protect sensitive information when its form, context and path through the system change?
Sensitive information can appear in prompts, retrieved content, memory and generated responses. We examine how privacy controls recognise it, where information may be exposed and how to evaluate protection without using real personal data in test cases.
Research areas
Evaluate detection of personal and sensitive information under changes to form and context.
Study information exposure across prompts, retrieval, memory and generated responses.
Explore data minimisation and access boundaries across an AI workflow.
Compare privacy protection, false positives and useful behaviour with controlled evaluation data.
Publications & related resources
AdvPIIBench - The Adversarial Collapse of PII Detection in LLM Interactions
USENIX Security 2026 · Poster session
From AdvPIIBench to the OWASP AI Testing Guide
AdvPIIBench: The Adversarial Collapse of PII Detection in LLM Interactions
SECAI 2026 @ ESORICS · Paper and oral presentation
Contribute to this research.
Explore student research, internships and collaboration with the lab.