Adversarial Resilience in Agentic AI Systems
An applied security study examining how tool-using AI agents behave when instructions, retrieved content, and connected services contain adversarial input.

Browse ongoing, completed, and exploratory work across SJIT's academic domains. Projects may result in prototypes, frameworks, technical reports, teaching cases, datasets, or future publications.
An applied security study examining how tool-using AI agents behave when instructions, retrieved content, and connected services contain adversarial input.
A practical investigation into how organizations can identify cryptographic dependencies, prioritize migration risk, and prepare internet-facing systems for post-quantum transition.
A study of how technical security findings can be translated into decision-ready risk narratives for organizations without large dedicated security teams.
An exploratory research track connecting device-level characteristics, edge computing, and the security assumptions made by software operating above the hardware layer.
An interdisciplinary project studying how organizations can document, review, and govern automated decisions without reducing AI governance to policy checklists.