Modeling Cooperation

Modeling Cooperation was founded in 2019 at the AI Safety Camp by Jonas Emanuel Muller, Vasily Kuznetsov, and Miles Tidmarsh, united by the goal of reducing risks from competition for the development of transformative AI. The project has since grown into a community of mission-aligned freelancers and volunteers who conduct game-theoretic analysis and computational modeling research. The organization investigates the dynamics of competition for transformative AI through two complementary approaches: game-theoretic analysis and computational modeling. Their research integrates insights from game theory, economics, technological races, and AI safety literature. A key research contribution is their work extending the Racing to the Precipice model by Armstrong, Bostrom, and Shulman, including analysis of how uncertainty and differing marginal returns from safety investments influence the risk of disaster. They have also published research evaluating the Windfall Clause policy within AI competition frameworks. Modeling Cooperation has developed several interactive software tools including the Safety-Performance Tradeoff (SPT) Model web app (built in collaboration with Robert Trager, Paolo Bova, Nicholas Emery-Xu, Eoghan Stafford, and Allan Dafoe), the Baseline web app for dynamic AI competition modeling, and a Racing to the Precipice web app. They also collaborate with Technology Strategy Roleplay on Intelligence Rising, a strategic simulation workshop that educates decision-makers about the dangers of AI races. The team includes co-founders Jonas Emanuel Muller (full-time since 2019, former software engineer based in Switzerland), Vasily Kuznetsov (mathematician and software engineer, former UN climate change secretariat worker), and Miles Tidmarsh (economics PhD candidate, former research economist for Australian Productivity Commission). Other team members include Paolo Bova (computer science PhD candidate working on policy impact simulations), Ben Harack (DPhil candidate at Oxford affiliated with Centre for the Governance of AI), Tanja Ruegg (project manager), and Jasmine Brazilek (full stack developer). Modeling Cooperation has received grants from Jaan Tallinn via the Survival and Flourishing Fund ($74,000 in 2020, $83,000 in 2022), the Center on Long-Term Risk Fund ($20,000), and the Foresight Institute ($77,116 in 2024). In 2025, they received a $66,000 SFF S-Process recommendation through the Freedom Track, with an additional $26,000 matching pledge available through September 2026. The organization is fiscally sponsored by Convergence Analysis.