Overview

An elite course in AGI Governance, explored through deep dives on pressing questions at the frontiers of the AI transition. After an initial level-setting session to shore up gaps in knowledge of AI itself and of the theory and practice of governance, we will lay foundations for and explore in depth a selection of topics such as:

1. frontier AI legislation at the state and federal level;
2. building societal resilience in the face of rapidly accelerating capabilities growth;
3. imbuing democratic practices into AGI alignment;
4. designing new governing institutions for the age of AGI and beyond, and
5. navigating fundamental changes in humanity’s self-understanding due to AGI’s advent.

By assembling a group of ambitious, early career researchers and practitioners who are jointly focused on steering society through the AI transition, we are aiming to build networks that will, by enduring, support cross-sectoral collaboration through the turbulent period ahead.

The fellowship will run for three weeks in Washington DC, in person only, from Tuesday, September 8, through Friday, September 25th. Travel to DC, as well as accommodation and meals/incidentals, will be funded by the fellowship. Fellows will engage in a full 5 day program, each week, of faculty lectures, seminars, collaborative group work, and engagement with guest speakers. Guests will be drawn from frontier labs, government, academia and civil society. The program will conclude with a set of group projects focused on materially contributing to AGI governance—illustrative examples would include prototype civic agents to support democratic decision-making; model legislation to address challenges posed by recursive self-improvement of AI systems; novel evaluations of LLM capabilities; a schema for using AI agents to advance AGI governance itself.

The program will be intentionally demanding, requiring significant pre-reading and a full-time commitment from fellows.

Who is eligible?

Applicants must have either a postgraduate degree in a relevant field or equivalent practical experience. They should be early career—typically within 8 years of receiving their undergraduate degree. They should already have decent foundations in technical understanding of frontier AI systems, and/or the governance of AI. This is the first year of this fellowship; our principal criterion will be excellence, therefore if in doubt, apply.

We are aiming for a cohort of 20-25 fellows. This is a rolling application, but apply by July 15 to be assured of full consideration.