MITRE Launches AI Incident Sharing Initiative with Industry Collaborators

MITRE’s Center for Threat-Informed Defense has made a groundbreaking announcement this week with the launch of the AI Incident Sharing initiative. This initiative is a collaborative effort involving more than 15 companies aimed at increasing community knowledge of threats and defenses for AI-enabled systems.

Under the umbrella of the center’s Secure AI project, the AI Incident Sharing initiative is designed to facilitate quick and secure collaboration on threats, attacks, and accidents involving AI-enabled systems. By leveraging the MITRE ATLAS community knowledge base, which has been collecting and analyzing anonymized incident data for the past two years, this initiative will provide a platform for a community of collaborators to access protected and anonymized data on real-world AI incidents.

Organizations can submit incidents via the web at https://ai-incidents.mitre.org/, with the opportunity to be considered for membership. The goal is to enable data-driven risk intelligence and analysis at scale, ultimately enhancing the collective defense of AI systems and mitigating external harms.

In addition to the incident sharing initiative, Secure AI has expanded the ATLAS threat framework to include information on the generative AI-enabled system threat landscape. This update incorporates new generative AI-focused case studies, attack techniques, and mitigation methods. The collaboration with Microsoft in November 2023 resulted in updates to the ATLAS knowledge base focused on generative AI.

Douglas Robbins, vice president of MITRE Labs, emphasized the importance of standardized and rapid information sharing to improve the defense of AI systems. This initiative is reminiscent of MITRE’s successful information-sharing public-private partnership with the Aviation Safety Information Analysis and Sharing database, which aims to prevent hazards in aviation. With representatives from industries such as financial services, technology, and healthcare, collaborators on Secure AI include major players like Microsoft, Intel, and JPMorgan Chase Bank.