Why this matters
AI competition is a collective action problem
The competition to build transformative AI is accelerating, with very little coordination on safety. Competitive pressure can make cutting corners on safety rational. This creates a collective action problem: a lab that slows down risks being overtaken, and a government that raises standards risks falling behind rivals that do not.
Modeling Cooperation makes those trade-offs more explicit through software and research for decision-makers, researchers, and educators working on AI governance.
What we do
Three ways we support AI governance
We help AI governance researchers and policymakers explore AI competition dynamics, test assumptions, and reason through trade-offs.
Scenario simulation platform
We build software that helps researchers and policy educators run structured simulations of AI competition and explore strategic trade-offs more directly.
See platform featuresInteractive models
Our interactive tools let researchers and policymakers test AI competition models, compare assumptions, and see how safety dynamics change across scenarios.
Explore our toolsAI competition research
We use game theory and agent-based simulation to study when AI developers invest in safety, when they do not, and which interventions may withstand competitive pressure.
Read our researchFlagship software
A scenario simulation platform for strategic workshops
Our platform supports real-time coordination, game mechanics, and workshop administration for structured policy simulations. It currently powers the Intelligence Rising workshops, run by Technology Strategy Roleplay.
About
Software and research since 2019
Modeling Cooperation is an independent group that started collaborating following the AI Safety Camp. We are a small team of software engineers and researchers working at the convergence of AI governance, software engineering, and quantitative modeling.
Our supporters include
- Convergence Analysis
- Survival and Flourishing Fund
- Foresight Institute
- EAF Fund
News
Read about our work
We post news about the grants we receive, the software we release, the research we publish, and our progress reports.
$108,000 from the Survival and Flourishing Fund's 2026 round
We're grateful to have received $108,000 in general support through the Survival and Flourishing Fund's 2026 round. It's the sixth SFF round to fund us, and the first in which Dustin Moskovitz funded our work, alongside Jaan Tallinn. The grant supports our ongoing development of the scenario simulation platform and our research on AI competition dynamics.
See SFF's recommendations$26,000 raised on Manifund, matching pledge fulfilled
Thanks to a generous donor on Manifund, we raised $26,000 and fulfilled the Survival and Flourishing Fund's matching pledge. SFF will match the donation dollar-for-dollar as general support. Together, the $52,000 will help us continue building our scenario simulation platform and conducting our research on AI competition dynamics.
View on Manifund$10,000 rapid grant from BlueDot Impact
We thank BlueDot Impact for awarding us a $10,000 rapid grant. The grant will cover the LLM API costs of developing AI-controlled players for our scenario simulation platform. These players will take the roles of states and AI companies in workshops. The aim is to help facilitators steer the game and participants learn more from it.
About BlueDot's rapid grants