Arcadia Impact AI Governance Taskforce Summer 2026 cohort—applications closing 10th May

🚨 AI Governance Taskforce Summer 2026 applications are open – deadline 10th May.

For senior or mid-career professionals transitioning into AI governance, focused on mitigating catastrophic risk.

This 12-week remote part-time opportunity is designed for experienced professionals who want to apply existing domain expertise to frontier AI governance, addressing acute and systemic catastrophic risks. You will work in small, high-calibre, multidisciplinary research teams producing policy-relevant research, led by Research Team Leaders and guided by expert practitioners based at respected organisations.

Projects and experts:

🔹 Building anticipatory capacity: frontier AI risk scenario frameworks for government
Elizabeth Dearden-Williams (UK AISI)

🔹 A playbook for corporate management of frontier AI incident knowledge under legal constraint
Sean McGregor (AI Incident Database)

🔹 Opportunities for advancing red lines by country – middle powers legislation
Caio Machado (The Future Society)

🔹 No treaty or weak treaty: Trade-offs between pathways to an international AI governance agreement
José Villalobos (The Future of Life Institute)

🔹 Testing and Evaluation of Agentic AI Systems in Military Command and Control
Heather Frase (Virginia Tech National Security Institute)

🔹 ‘If Anyone Builds It, Everyone Needs to Understand It’ — Understanding as a product of frontier AI safety review – building shared comprehension for collective governance
Professor Robin Bloomfield (City University, University of London)

What this involves:
👥 Applied research in small teams
📄 Policy-relevant outputs (papers, briefs)
🌍 Fully remote, global, part-time (8 hrs/​week), 12 weeks
💼 Selective, unpaid research fellowship
📅 Deadline: Sun 10 May 2026
📅 Cohort dates: 1 June − 21 August 2026
🔗 More info and apply: https://​​www.arcadiaimpact.org/​​ai-governance-taskforce

For queries, please email:
Ben R Smith, Taskforce Lead
ben [at] arcadiaimpact [dot] org