During my recent ten-month sabbatical at the Council on Foreign Relations (CFR), I served as the Technologist in Residence, exploring the intersection of AI, warfare, and national policy. This was a significant period of professional growth, as I broadened my perspective from a purely technical one to understand how technology is adopted and shaped by policy within the U.S. government.
My work focused on three core areas.
First, my research, co-authored with Jonathan Panter, examined the increasing use of AI-powered autonomous weapons and their misalignment with governing policies. We argued that the belief that humans can effectively control these systems during wartime is a dangerous illusion that could disadvantage democratic nations against less scrupulous adversaries. Instead, it is crucial for governments to integrate human input and values into these systems now, before conflicts begin, to avoid poor choices under duress and ultimately save lives. This topic was explored in our article in Foreign Affairs, titled "AI Weapons and the Dangerous Illusion of Human Control."
Second, I investigated the role of Large Language Models (LLMs) in high-level strategic decision-making. While some have raised concerns about LLMs escalating conflicts, my findings indicate that simple, non-technical interventions can effectively manage these risks, suggesting a more measured outlook on their potential for driving escalation.
Finally, my co-authored work with Adam Segal argued for a more realistic approach to the U.S.'s global position in AI. Rather than solely striving for dominance, American policymakers should prepare for a future of shared AI leadership. Failure to adapt to and leverage advancements from other countries, like China, would be a more significant risk than simply falling behind. This was the focus of our article in Foreign Affairs, titled "What If China Wins the AI Race?"
Beyond my research, the sabbatical gave be a broader understanding of how computing technology is adopted, its limits, and its implications in a geopolitical context. I learned how to write papers that analyze the implications of technology within a geopolitical framework. I expanded my network beyond the computer science community to include thought leaders in national security, defense, and policy. I also had the opportunity to organize roundtables with leaders from the Department of Defense and industry, and brief members of Congress on these critical issues. This experience has significantly broadened my ability to contribute to the national conversation on technology and geopolitics
A special thank you to my colleagues and co-authors Jonathan Panter, Adam Segal, and Kat Duffy, my RA Max Hippold, my roundtables co-organizer Nick Beim, and the rest of the CFR fellows and staff members who made this a memorable experience.