# Theory of change
<div class="pills-container">
<span class="pill">Last Updated: December 31, 2025</span>
</div>
### Fundamental impact
Increase epistemic resilience against AI x-risks.
### Intermediate outcome
Institutions outside frontier AI hubs can understand AI risks, set their own priorities, and act on them.
### Outputs
1. Recognized roles and mandates for AI safety and risk oversight inside governments, civil society, and other decision-making institutions outside frontier hubs
2. Cheap and usable AI safety tools that institutions can run themselves and use to coordinate with others.
### Activities
1. Help people outside frontier hubs do real AI safety-relevant work while creating or expanding roles that give them decision influence.
2. Translate AI safety research into operational infrastructure that enables non-frontier actors to coordinate, escalate issues, and make binding decisions.