James Sykes is an Advisor for AI Risk Management for the Recursive Self Improvement Initiative at the Institute for Security and Technology (IST). He is a PhD student in Statistics at the University of Warwick, where he develops methods to study the stochastic dynamics of multi-agent AI systems, and holds a Master of Mathematics from Durham University. Alongside his doctoral research, James works on quantitative AI risk modeling. He has contributed to UC Berkeley Center for Long-Term Cybersecurity (CLTC) work on using Bayesian networks to operationalize risk thresholds, and then continued this line of research with researchers at SaferAI as part of the SPAR program, focusing on expert elicitation for AI risk models. He is a co-author of “Open Problems in AI Risk Modeling: Insights from a Workshop on the Technical Foundations of AI Risk Modeling.”