AI Readiness Institutes: how small states can prepare for fast AI progress
Most small European states have no AI Security Institute. We propose AI Readiness Institutes: like an AI Security Institute, they track the capabilities and risks of frontier models; they also help public bodies, critical infrastructure and researchers keep access to top models and compute, and prepare for very fast AI progress. Both jobs need the same expertise and contacts, and the problems tend to arrive together. Matthias Samwald and I are working on establishing one in Austria, and on making the idea replicable elsewhere.
Get in touch
- Working on something similar in your country?
- Know people who should be part of this (cyber, evaluations, AI Security Institute experience)?
- jason.hoelscherobermaier@accelerate-europe.org 路 LinkedIn
Sources
- Map, as of October 2026. Running: UK, France (INESIA), Germany (AISI Deutschland). Planned: Spain (AI Security Institute within AESIA, Plan IA360, measure 4). Proposed, not yet an institute: Norway (AI security mandate and evaluation capacity for KI Norge, 2027 budget proposal)
- June 2026 access cut-off: Anthropic, statement on the US directive to suspend access to Fable 5 and Mythos 5, 12 June 2026
Further reading
- Woodward and Rogoyski, Sovereign by necessity? Frontier AI export controls, cyber security, and the limits of national AI capability (2026)
- GovAI, Delays to frontier AI in the EU and UK
- Markus Anderljung and Stephen Clare, What can middle powers do for frontier AI governance?
- IGC, How to build a high-impact AI team inside the state
The initiatives
- Accelerate Europe: a European agenda for the age of transformative AI
- Transformative KI 脰sterreich: a campaign to prepare Austria for transformative AI