GovAI relaunches as a nonprofit, with plans to expand field-building work

What stays the same and what changes

According to GovAI, its mission and core research activities remain largely unchanged. What the new nonprofit structure changes is the room to maneuver: the organization says the greater flexibility will let it expand its field-building work.

The group lists its continuing activities as producing, supervising and coordinating research; running a twice-annual fellowship program for junior AI governance researchers; hosting internal research seminars and a public seminar series; advising decision-makers; and offering career advice, connections and informal mentoring to promising people entering the field.

For the coming year, GovAI plans to host its inaugural AI governance conference, begin awarding student research prizes, and explore an expansion of its policy-advising work.

A two-year period of experimentation

GovAI says that over approximately the next two years it intends to experiment and learn more about how it can provide the most value in its current form. It states that it may adjust or halt current or planned activities if it is not sufficiently convinced of their impact.

The organization also says it may trial additional activities, such as grantmaking and scholarship programs, or a program reminiscent of the Forethought Fellows program.

Leadership, team and funding

GovAI is based in Oxford, in the same building as the Future of Humanity Institute, the Global Priorities Institute and the Centre for Effective Altruism, and works with a global network of collaborators and affiliates across several institutions.

Acting Director Ben Garfinkel leads the organization. He is a Research Fellow at the Future of Humanity Institute and head of its AI Governance Team, and has been involved with GovAI and its predecessor organizations for five years. President Allan Dafoe, the founder and previous Director of GovAI, advises the organization and collaborates on research projects. He currently heads DeepMind's Long-Term AI Strategy and Governance Team.

The rest of the core team includes:

• Markus Anderljung, Head of Policy and Research Fellow
• Joslyn Barnhart, Applied Research Lead
• Alexis Carlier, Head of Strategy
• Noemi Dreksler, Survey Researcher
• Anton Korinek, Economics of AI Lead
• Anne le Roux, Operations Manager
• Robert Trager, Strategic Modeling Team Lead
• Eoghan Stafford, Strategic Modeling Researcher

Toby Shevlane will soon join as a Research Fellow. The Advisory Board consists of Ajeya Cotra, Allan Dafoe, Helen Toner, Tasha McCauley and Toby Ord. GovAI thanks Open Philanthropy, the Centre for Emerging Risk Research and Effective Altruism Funds for financial support, and the Centre for Effective Altruism for temporary fiscal sponsorship.

Open roles and the Summer Fellowship

GovAI is hiring a Chief of Staff and a Research Fellow. The Chief of Staff would act as the "central node" within GovAI, reporting only to the Director, and would take on a range of responsibilities that mostly sit with the Director today. The organization believes the right candidate could significantly increase its long-run impact and ability to expand its activities.

Research Fellows are expected to produce research on important problems and open questions in AI governance. GovAI says it is interested in candidates from a range of disciplines who have shown an ability to produce excellent research and care deeply about the long-run impacts of AI. The role offers significant research freedom, access to a broad network of experts and opportunities for collaboration.

Applications are also open for the Summer Fellowship Program, which lets early-career individuals spend three months on an AI governance research project, learn about the field and explore ways to contribute.

What it means for the field

For people interested in AI governance, the relaunch points to more entry routes: open hiring, a fellowship for early-career researchers, planned student research prizes and a first conference. GovAI's own caveat is that its plans are provisional, since it may change or stop activities that do not prove their impact during the coming two years.