Regina Garcia
A partnership between Mitre and Microsoft has integrated multiple open source tools to create a plug-in aimed at helping security professionals discover vulnerabilities in artificial intelligence and machine learning platforms.
The team produced Arsenal using Microsoft’s Counterfit security test automation tool along with Mitre’s Adversarial Threat Landscape for Artificial-Intelligence Systems and Caldera cybersecurity framework, the nonprofit company said Thursday.
Counterfit is intended to facilitate attack emulation with ML-powered systems, while the Caldera platform allows the user to build and automate adversary profiles.
Mitre’s ATLAS database is based on real-world studies of adversarial tactics and techniques and ML case studies.
Charles Clancy, senior vice president and general manager of MITRE Labs, said ML workflow complexity could affect the process of identifying system vulnerabilities.
He added that specialists could employ the Arsenal-Caldera integration in efforts “to discover novel vulnerabilities within the building blocks of an end-to-end ML workflow and develop countermeasures and controls to prevent exploitation of ML systems deployed in the real world.”
The two companies plan to update these tools with additional adversary profiles and security methods based on threat research documentation.
