Chandler Smith

About Me

My name is Chandler, and I am currently a MATS scholar collaborating with Jesse Clifton, researching multi-agent systems. I also am a MARS participant, where I am part of a team lead by Rhys Ward, looking into model deception and intention. In addition to AI Safety research, I work as a Technical Architect Fellow at IQT where I help support the national security and intelligence community. Previously, I worked as an engineer for Dimagi on global health and COVID response projects. On a personal note, I am based out of Washington DC where I live with my partner, cat, and dog. I enjoy running, going out to eat, and spending time with friends.

Research Interests

I am particularly interested by both the risks posed by multi-agent solutions and the potial use of agents for oversight. Promising research areas include Scalable Oversight, Cooperative AI, and assessing individual model traits such as deception and escalation.

Publications and Ongoing Work

EscalAtion: Assessing Multi-Agent Risks in Military Contexts

Our project investigates the potential risks and implications of integrating multiple autonomous AI agents within national defense strategies, exploring whether these agents tend to escalate or deescalate conflict situations. Through a simulation that models real-world international relations scenarios, our preliminary results indicate that AI models exhibit a tendency to escalate conflicts, posing a significant threat to maintaining peace and preventing uncontrollable military confrontations. The experiment and subsequent evaluations are designed to reflect established international relations theories and frameworks, aiming to understand the implications of autonomous decision-making in military contexts comprehensively and unbiasedly. Aurhors: Gabriel Mukobi*, Anka Reuel*, Juan-Pablo Rivera*, Chandler Smith*