CISA leads first tabletop exercise for AI cybersecurity

CISA led a four-hour AI tabletop exercise under the Joint Cyber Defense Collaborative with 50 experts from 15 companies (including Microsoft, OpenAI, IBM, Palantir, Cisco, Palo Alto Networks) and U.S. and allied cyber agencies to define AI-related incident categories, information-sharing needs, and public‑private coordination. The exercise supports an AI security incident collaboration playbook planned for late 2024 and complements recent DHS/CISA initiatives such as AI security guidance for critical infrastructure, an AI safety and security board, and a supply-chain security working group.