Scientists and Strategists - National Security AI in Practice: Evaluation, Governance, Human-Machine Integration, and Other Emerging Issues

As artificial intelligence becomes increasingly embedded in national security, the most consequential challenges extend well beyond the performance of individual models. This talk will examine three interconnected priorities for national security AI: evaluating AI-enabled systems under operationally realistic conditions; evolving governance for rapidly evolving capabilities; and optimizing human-machine integration as AI becomes an increasingly important part of decision-making. Drawing on developments in AI decision-support systems, agentic AI, and embodied systems, the discussion will also examine emerging issues shaping the next phase of national security AI and consider AI as a social-cultural technology—one whose effects depend as much on people, culture, organizations, doctrine, and institutional incentives as on chips, data, models, and compute. This dialogue will be moderated by CMIST Director Audrey Kurth Cronin. Doors open at 4:30 p.m.; refreshments served * Seating in the Grand Room is reserved for registered guests on a first-come, first-served basis. An overflow room will be available for viewing, also first-come, first-served. Once registration is full, walk-ins will be directed to the overflow room. Late-arriving registered guests may be directed to overflow seating.