Comp-HuSim: Complex Human Simulations

Comp-HuSim: Complex Human Simulations

Michael G. Yankoski, William & Mary and Colby College (Affiliate) 
Trenton W. Ford, William & Mary

How can modern Large Language Models (LLMs) be used to generate complex
pseudo-personalities in social simulation systems? Will this approach be effective? Will these systems degrade over time? Will they be capable of simulating and modeling emergent social and political behaviors? How might we effectively model personality and individual history in the software stack and prompts? How might the utilization and deployment of complex human simulations at scale be utilized in psychological operations?

Dr. Michael G. Yankoski is a research scientist in the Data Science Program at the College of William & Mary. He recently completed postdocs at the Davis Institute for Artificial Intelligence at Colby College, as well as the Department of Computer Science and Engineering at the University of Notre Dame. His PhD is in International Peace Studies and Ethics from the University of Notre Dame's Kroc Institute for International Peace Studies.

Dr. Trenton W. Ford is an Assistant Professor of Data Science at William & Mary whose research focuses on the intersection between Artificial Intelligence, Machine Learning, and Society. Specifically, his lab uses AI and ML to learn about, and potentially predict, human behavior. Dr. Ford is currently working on questions of misinformation and democracy while also working on projects to model human cognition to enable high fidelity individual simulations with an end goal of community level simulations.