FIG runs a part-time, remote-first 12-week research fellowship connecting early-to-mid-career researchers with experienced project leads working on AI safety, AI governance, and AI sentience.
FIG runs a part-time, remote-first 12-week research fellowship connecting early-to-mid-career researchers with experienced project leads working on AI safety, AI governance, and AI sentience.
People
Updated 05/18/26AI Research Fellow
Research Fellow
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26Future Impact Group (FIG) is a fellowship organization founded in 2023 and based in Oxford, UK. FIG was co-founded by Suryansh Mehta and Callum Evans with the goal of making high-impact research on pressing global problems more accessible to talented people who cannot take on traditional full-time roles. The organization's managing director is Luke Dawes, a former teaching fellow at BlueDot Impact and diplomat, and its programme operations are led by Marta Krzeminska, who previously supported AI safety research at the Centre for Long-Term Resilience. FIG's flagship program is a 12-week, part-time, remote-first fellowship in which participants serve as research associates on specific projects supervised by experienced leads. Fellows commit 8 or more hours per week and work across three core areas: AI Policy (covering national and international governance, economic effects of advanced AI, and technical evaluations), Philosophy for Safe AI (combining technical AI safety work on topics like LLM interpretability and reward-seeking behavior with foundational philosophical questions), and AI Sentience (investigating consciousness models, digital minds, and AI welfare governance). Project leads affiliated with FIG have come from a wide range of high-profile organizations including the University of Oxford, the Centre for the Governance of AI, Google DeepMind, Anthropic, MIT, the London School of Economics, the Global Priorities Institute, METR, and Polaris Ventures. In the Winter 2025 cohort, the program hosted 28 project leads, 33 projects, and 62 fellows. FIG has also run a special collaboration with AI researcher Yoshua Bengio, in which fellows co-authored policy reports on topics such as insurance and liability as levers for AI safety, AI-driven concentration of power, and military AI. FIG provides fellows with co-working sessions, career guidance, issue troubleshooting, opening and closing events, research sprints, and guest speakers. The fellowship is typically unpaid but FIG maintains limited funding for applicants facing financial hardship, and some project leads or their institutions provide compensation. Toward the end of each cohort, FIG evaluates projects and may offer limited funding to extend particularly high-impact work. FIG is funded by Open Philanthropy and Longview Philanthropy. Alumni of the program have received offers from organizations including OpenAI, the UK Department for Science & Technology, the UN Envoy on Technology, and the Centre for the Governance of AI. Research outputs have included blog posts for think tanks, papers under consideration at NeurIPS and ICML, and confidential policy memos to support meetings with cloud compute providers.
Theory of Change
Updated 05/18/26FIG's theory of change holds that many talented researchers who could contribute to reducing AI-related risks are blocked from doing so by structural barriers: the need for full-time employment, geographic constraints, or lack of access to established networks. By offering part-time, remote-first research roles with experienced supervisors and a supportive community, FIG enables these researchers to develop expertise, build credentials, and produce useful research outputs. In the near term this generates direct research value (policy memos, technical papers, governance analyses). In the medium term it creates a pipeline of trained, credentialed researchers who enter high-impact organizations in AI governance and safety. The cumulative effect is a larger and more capable field working on the problems most likely to determine whether transformative AI goes well.
Grants Received– no grants recorded
Updated 05/18/26Projects
Updated 05/18/26A 12‑week, part‑time and remote‑first Winter 2025 research fellowship where associates work at least 8 hours per week on projects in AI governance, technical AI safety, and AI sentience under experienced project leads.
Discussion
No comments yet. Be the first to share your thoughts.