Project Silicon: Call for Participants

[Stephen Leider]

Dear Colleagues,

 

We are launching Project Silicon, a large-scale collaborative research initiative - developed in coordination with Management Science in the registered-report format - designed to rigorously assess whether today's leading AI models can accurately simulate human decision-making in behavioral experiments.  The goal is to evaluate the usefulness of these models as an experimental design tool, e.g. for pre-testing novel experimental designs.  We are assembling a diverse portfolio of new and unpublished behavioral experiments across economics and operations management, running AI simulations across the full spectrum of leading foundation models, and comparing the simulated results against actual human data. Because every experiment is unpublished, the AI models cannot simply retrieve known results from their training data-they must genuinely simulate human behavior.

 

We are seeking researchers planning behavioral experiments in operations management or economics to be completed by January 2027. Participation requires minimal effort beyond your normal research activities: you share your finalized experimental materials, we construct AI agents and run simulations across every leading foundation model, and you run your experiment as planned. Our team handles all aspects of the AI simulation-no technical expertise is required. The project is fully funded, at no cost to participants, and requires no IRB amendments.

 

In return, participants receive:

• Co-authorship on the Project Silicon paper
• Full AI simulation output for your specific experiment
• Simulation support for follow-up experiments and revisions
• Access to our open-source simulation toolkit

All experimental materials are kept strictly confidential. We are accepting registrations now through July 2026 - the full call for participants is attached, and you can register your interest at www.project-silicon.org or reach us at project...@umich.edu.

 

Best regards,

Stephen Leider, University of Michigan

Andrew Davis, Cornell University

Andrew Wu, University of Michigan

Jing Wu, Chinese University of Hong Kong