University of Oxford

The University of Oxford has no single founding date, with evidence of teaching dating to 1096, making it the oldest university in the English-speaking world and one of the oldest in continuous operation in the world. It is a collegiate research university comprising 43 colleges and a range of academic departments organized across four divisions. As of 2024, it employs approximately 16,905 staff and hosts 22,735 full-time students. Its total annual income was approximately £3.1 billion in 2023-24, supported by a combined endowment of £8.7 billion. Oxford has been one of the most important institutional homes for AI safety and existential risk research globally. The Future of Humanity Institute (FHI), founded in 2005 by philosopher Nick Bostrom as part of the Oxford Martin School, pioneered research on existential risk, AI alignment, AI governance, longtermism, and global catastrophic risk. FHI spawned numerous influential ideas and spin-off organizations, including the Centre for the Governance of AI (GovAI), which became an independent nonprofit in 2021. FHI was closed on 16 April 2024 after facing what it described as increasing administrative headwinds within the Faculty of Philosophy, with a hiring and fundraising freeze imposed from 2020 and non-renewal of staff contracts in late 2023. Following FHI's closure, AI safety and governance research at Oxford continues through several channels. The Oxford Martin AI Governance Initiative (AIGI), co-directed by Robert Trager and Michael Osborne and housed at the Oxford Martin School, focuses on understanding risks from advanced AI through rigorous technical and policy research across five domains: frontier AI governance, technical AI governance, international AI governance, social impact of emerging technologies, and China AI governance. AIGI offers novel DPhil studentships bridging advanced AI systems and governance. The Oxford Internet Institute runs a research programme on AI, government, and policy supported by the Dieter Schwarz Foundation. In April 2025, Oxford researchers were awarded grants through the UK Government's £59 million ARIA Safeguarded AI programme to develop safety-first AI approaches. The Oxford AI Safety Initiative (OAISI), a student-led community that became independent from the university in Summer 2024, continues running ML safety bootcamps and research projects.