3 papers
cs.CR2026
CorrectGuard: Eyes-Off Correctness Estimation for Black-Box Security Guardrails
Adam Faulkner, Nil-Jana Akpinar, Matthew Dressman
AI services increasingly rely on black-box security guardrails, yet privacy-preserving model auditing regimes often cannot measure how well these systems perform in both a human ey…
cs.AI2026
Magnet: Detecting Cross-Session AI Misuse Through Capability Accumulation
Natalie Isak, Matthew Dressman
The most capable AI deployments are not single models but ensembles of specialized agents that delegate and act in coordination. This architecture unlocks powerful new capabilities…
cs.CR2025
BinaryShield: Cross-Service Threat Intelligence in LLM Services using Privacy-Preserving Fingerprints
Waris Gill, Natalie Isak, Matthew Dressman
The widespread deployment of LLMs across enterprise services has created a critical security blind spot. Organizations operate multiple LLM services handling billions of queries da…