Directory

Impact Research Groups (IRG) is a structured research training programme operated by Arcadia Impact, a charitable organisation registered in England and Wales (Charity No. 1212323). The programme is designed to help talented and ambitious students take their first steps toward careers in high-impact research by working in small, mentored teams on real research questions.

Over 8 weeks, participants explore a research question within one of five key focus areas: technical AI safety, AI governance, biosecurity, global health, or animal welfare. The programme culminates in a competitive review, with a £2,000 prize awarded to the best overall project. The kick-off weekend requires in-person attendance in London, while the remainder of the programme can be completed remotely. Cohorts have included approximately 40 participants…

IRG's theory of change centers on talent pipeline development: by providing structured, mentored research experience to promising students early in their careers, the programme builds a generation of researchers equipped to work on the world's most pressing problems including AI safety and biosecurity. Participants gain hands-on research skills, exposure to the effective altruism and AI safety research communities, and concrete outputs they can use to launch careers at high-impact organisations. This supply-side intervention in the AI safety talent ecosystem is intended to compound over time as alumni go on to work at leading AI safety labs, policy bodies, and research organisations.

Safe AI London (SAIL) is a project of Arcadia Impact, a UK-registered charity (Charity Number: 1212323) that enables individuals in London to use their careers to tackle the world's most pressing problems. SAIL focuses specifically on AI safety awareness and career development, aiming to raise awareness of risks from advanced AI particularly within universities, and to provide high-quality resources and support to students and professionals interested in contributing to AI safety.

SAIL's main programs include an 8-week AI Safety Course offered in collaboration with AI Safety Collab, with two tracks: an Alignment Track (based on AI Safety Atlas curriculum) and a Governance Track (based on BlueDot Impact curriculum). Discussion sessions are primarily online with in-person events for London participants. SAIL also runs the Research Engineers Club, which uses pair-programming to develop research engineering skills through paper replications, with support from expert speakers and in-person engineering sprints at the London Initiative for Safe AI (LISA) coworking space…

SAIL aims to expand the pipeline of talent working on AI safety by raising awareness at London universities and providing accessible entry points into the field. By offering structured courses, research engineering programs, and community events, SAIL helps individuals identify AI safety as a priority cause area and develop the skills needed to contribute effectively. The underlying assumption is that talent and community bottlenecks are a key constraint on AI safety progress, and that supporting early-career individuals to enter the field compounds over time into greater research and governance capacity.

The Orion AI Governance Initiative is a project of Arcadia Impact, a registered charity in England and Wales (charity number 1212323) founded in 2022 and based in London. Arcadia Impact operates several talent development programmes aimed at addressing global catastrophic risks, including AI safety and AI governance. The Orion initiative was launched in 2024 specifically to address a perceived talent gap in the AI governance field by supporting students and early-career professionals to transition into AI policy careers.

The programme has three flagship components. The AI Policy Accelerator is a 9-week part-time programme that first provides a foundational curriculum covering AI fundamentals, hardware, supply chain considerations, and governance frameworks, followed by a 4-week supervised research project phase in which participants work in small teams to produce substantive research outputs on real AI governance questions. The Mentorship Programme pairs participants with working AI governance professionals, including those from policy teams at frontier AI laboratories and prominent think tanks such as the Centre for the Governance of AI, RAND, and the Ada Lovelace Institute. The…

The Orion AI Governance Initiative operates on the premise that the AI governance field faces a critical talent shortage, and that the trajectory of AI policy will depend significantly on who enters and shapes the field in the near term. By providing students and early-career professionals with technical literacy, hands-on policy research experience, mentorship from established practitioners, and direct placements into governance roles, Orion aims to accelerate the career development of high-potential individuals and increase the number of skilled people working on AI governance. The causal chain runs from participant recruitment and training through skill development and professional network formation to job placement in high-leverage AI policy positions, ultimately increasing the quality and quantity of expert human capital available to shape beneficial AI governance outcomes.