CENTER FOR HUMAN-COMPUTER
INTERACTION + DESIGN

• Christopher Schuh, Dean of McCormick School of Engineering
• Darren Gergle, Co-Director of the Center for Human-Computer Interaction + Design
• Jessica Hullman, Ginni Rometty Professor of Computer Science and Faculty Fellow at the Institute for Policy Research at Northwestern University

Lightning Talks and Q&A

• The Mixed Subjects Design: Treating Large Language Models as Potentially Informative Observations - David Broska
• Why Human Interaction Matters for AI Evaluation—and How User Simulators Can Help - Serena Chang
• Valid Survey Simulations with Limited Human Data - Kristina Gligoric

Lightning Talks and Q&A

• Large language models that replace human participants can harmfully misportray and flatten identity groups - Angelina Wang
• When Can We Trust Experiments on Digital Twins?  A Potential Outcomes Framework for Causal Inference with LLM Simulations - Patryk Perkowski
• Causal Inference in Experiments with Mixed-Subjects Designs - Austin van Loon

• Semantic Structure in LLM Internal Representations - Austin Kozlowski
• Detailed Self-Reports Can Reduce Bias in LLM-Simulated Survey Respondents - Jonne Kramphorst
• Synthetic Consumers in Practice: Methods, Validation, and Applied Challenges - Michael Spadafore
• Social Science and the Bitter Lesson: Thoughts on Valid Social Science when Machines Do Research - John Horton