We picture the compact as a hypersphere of intersecting formal and informal agreements among firms, governments, workers, researchers, communities, and the wider public, as well as the processes to adapt them over time. We envision a set of revisable agreements operating across multiple scales, with room for experimentation and learning.
Effective governance depends on testing interventions before scaling. Methods such as randomized trials and quasi-experimental designs require comparisons and evidence that often do not exist for current policy proposals. We want to build alternative methods, including several that use AI itself, and explore new mechanisms for sharing and producing data.
See work on evidence
Standard-setting bodies, scientific assessment panels, and shareholder trusts each have distinct functions and advantages. Future forms of organization should be designed around the capability gaps our existing institutions miss, including speed, verifiability, and technical oversight.
See work on institutions
Firm and regulatory approaches are increasingly converging. Plurality provides a check on concentrated power and keeps room open for a more diverse, resilient technology ecosystem.
See work on plurality
As technology systems become more complex, the public needs better ways to understand who holds power, how change happens, and where an intervention might work.
See tools and maps
Frontier technology governance requires coordination, while relevant expertise is often scattered across fields and communities. We build working relationships that let those perspectives shape a shared strategy.
See partnership work