Equilibria Network is a collective intelligence research organization studying how coordination mechanisms affect group outcomes, with a focus on multi-agent AI safety and democratic resilience.
Equilibria Network is a collective intelligence research organization studying how coordination mechanisms affect group outcomes, with a focus on multi-agent AI safety and democratic resilience.
People
Updated 05/18/26Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26Equilibria Network is a collective intelligence research organization working on how groups — human, AI, or mixed — coordinate and make better decisions. The organization explores the design space of coordination mechanisms through simulations and rigorous mathematical foundations, treating markets, democracies, and other coordination structures as comparable formal systems. Their research agenda is structured as a multi-phase roadmap progressing from theoretical foundations (developing a unified mathematical language for coordination and a taxonomy of agents) through infrastructure construction, validation, and ultimately the design of novel collective intelligence forms. Current outputs include published work on system-level safety evaluations, agent phylogenies, and spectral signatures of gradual disempowerment. Equilibria Network builds two primary tools: the Democratic Resilience Lab, a governance simulation framework for stress-testing voting systems and governance structures against adversarial scenarios; and the DeSci Bridging Lab, a research discovery environment built on open protocols. The organization partners with Digital Democracy World on liquid futarchy implementations. The core team includes Jonas Hallgren (Co-Director and AI safety researcher focused on multi-agent coordination), Aaron Halpern (Strategy and Ecosystem Lead with a PhD from University College London on the origins of genetic coding), and Markov Grey (Technology and Communication Lead, technical generalist and co-author of an early AI safety textbook). Advisors include researchers from Oxford University and the Cooperative AI Foundation. Jonas Hallgren is also a co-founder of AI Safety Sweden and former lead organizer of Effektiv Altruism Uppsala.
Theory of Change
Updated 05/18/26Equilibria Network believes that many AI risks are not about a single system going wrong, but about gradual degradation of societal coordination capacity. By building mathematical frameworks and simulation infrastructure to understand how groups of agents coordinate — and where those coordination mechanisms break down — they aim to give policymakers, AI labs, and safety researchers the tools to evaluate governance and deployment decisions before they are implemented at scale. Their causal chain runs from foundational theory and simulation tools, to practical adoption by institutions designing coordination systems, to AI deployment environments that are more robust against emergent failures and adversarial dynamics.
Grants Received– no grants recorded
Updated 05/18/26Projects
Updated 05/18/26An open-source, process-centric multi-agent simulation platform developed by Equilibria Network for exploring coordination mechanisms and collective intelligence dynamics.
Research agenda developing adversarial, multi-agent evaluations to stress-test societal defensive processes—such as democratic mechanisms, scientific consensus formation, and social dynamics—so as to measure system-level AI safety properties under coordinated attacks.
Discussion
No comments yet. Be the first to share your thoughts.