The RAND Center for the Geopolitics of Artificial General Intelligence (AGI) conducts “Day After” AGI exercises using RAND’s Infinite Potential platform to understand how the United States should respond to and prepare for potential artificial intelligence (AI) developments in the future.1 These exercises simulate a National Security Council Principals Committee (PC) convention to recommend a U.S. government response to developments in frontier AI. In each exercise, participants are presented a scenario that represents both (1) an acute crisis for U.S. national or economic security and (2) a signpost on a path to a transformative AI future. Facilitated by a simulated National Security Advisor, participants roleplay PC members diagnosing the implications of the scenario for U.S. national security, discussing courses of action presented by their “staffs” and recommending a path ahead for the President of the United States.
This report presents findings related to a scenario named Cyber Surprise, in which players confronted the sudden deployment of an advanced cyber-AI capability by the People’s Republic of China (PRC) to reinforce its own networks, and in the second turn, responded to a sudden large-scale cyberattack.
Center for the Geopolitics of Artificial General Intelligence
RAND Global and Emerging Risks is a division of RAND that delivers rigorous and objective public policy research on the most consequential challenges to civilization and global security. This work was undertaken by the division’s Center for the Geopolitics of Artificial General Intelligence (AGI), which is committed to helping decisionmakers understand, anticipate, and prepare to navigate the national security and geopolitical implications of AGI. For more information, visit http://www.rand.org/geopolitics-of-agi.
——————————————————————
The RAND Center for the Geopolitics of Artificial General Intelligence (AGI) conducts a series of “Day After” AGI exercises using the RAND Infinite Potential platform to support the center’s overall goal of understanding how the United States can respond to the technically credible but uncertain potential for developments in AGI.2
These exercises are designed to simulate a National Security Council Principals Committee (PC) convention to recommend a U.S. government response to developments in frontier AI.3 In each exercise, participants are presented a scenario that represents both (1) an acute crisis for U.S. national or economic security and (2) a signpost on a path to a transformative AI future. Specifically, these exercises aim to understand how the United States might wish to respond to sudden changes brought about by AI, and in particular what judgments participants seek to make, what capabilities they believe they need to have on hand in the scenario presented to them, and what playbooks they believe they need to develop to guide their responses.
This after-action report summarizes the results of six runs of the “Cyber Surprise” scenario, in which players confronted the sudden deployment of a powerful cyber-AI capability by the People’s Republic of China (PRC). These scenarios were run in spring and fall 2025. Across the exercises, two key themes unique to this scenario emerged in players’ discussions and decisionmaking:
- Participants recommended aggressive action when presented with advanced PRC cyber- AI capabilities. Participants responded to the PRC’s development of advanced cyber-AI that reduced the United States’ ability to access PRC cyber networks by recommending aggressive use of remaining cyber capabilities to acquire the Chinese cyber-AI model, disrupt its use, and/or pursue other cyber options to maintain intelligence-gathering capabilities. Participants expressed that the scenario presented the U.S. government with a use-it-or-lose-it situation regarding remaining cyber access, and they generally favored using what capabilities they had available for maximum advantage, even if such actions might be considered highly escalatory. Participants expressed concerns over the potential consequences of the PRC’s new cyber-AI capabilities for the long-term balance of power and sought to find ways to ameliorate that potential shift, more than they expressed concerns regarding potential PRC reprisal from a sudden U.S. escalation in cyberspace.
- Participants expressed uncertainty about the best means for engaging allies and partners, even if they agreed on the importance of engagement. In all iterations of this scenario,
2 These exercises have also been referred to as “Day After AGI” games. The Infinite Potential platform supports and administers these exercises, seeking to bring them to diverse stakeholders within and outside RAND to explore the policy challenges presented by the potential development of AGI.
3 These exercises draw inspiration from the “Day After” methodology developed by RAND. For a deeper discussion of this methodology, see Chapter 12 of Stuart Johnson, Martin C. Libicki, and Gregory F. Treverton, eds., New Challenges, New Tools for Defense Decisionmaking, RAND Corporation, MR-1576-RC, 2003.
v
participants raised questions on how to engage allies and partners. In particular, participants focused on what to reveal to allies and what to conceal from them regarding PRC and U.S. cyber capabilities. Participants agreed on the importance of having a high-quality engagement plan with allies and partners because they believed it necessary to work with other nations in response to a confrontation with the PRC concerning potential offensive use of cyber- enhancing AI.
Bankwatch 