The Icelandic AI Model
A proven framework for national AI capacity building. Designed, coordinated, and delivered by BTR; validated at national scale in Iceland; built to be adapted by any country or large institution.
Overview
In 2025, Iceland became the first country to deploy a comprehensive national AI capacity building program, starting with education: 700+ teachers applied from 49 of Iceland's 64 municipalities, gaining access to frontier AI tools from Anthropic and Google with structured training and support, roughly 10% of the nation's K-12 teaching workforce. Coordinated by BTR on behalf of Iceland's Ministry of Education and Children, the program moved from concept to classrooms in every region of the country in six months.
The Icelandic AI Model is the replicable framework that made this possible. It adapts to any country seeking to build AI capacity across its public workforce, starting with whichever sector offers the highest leverage: education, healthcare, public administration, or others.
Teachers applied
Of K-12 teachers nationwide
Concept to national scale
Countries with formal interest
How It Works
Five Phases
Each phase is designed to de-risk AI adoption while delivering measurable outcomes quickly.
Discovery
Identify high-impact use cases with government stakeholders. Map institutional readiness, regulatory landscape, and workforce needs.
Partnership
Broker agreements with frontier AI vendors who contribute technology, training, and deployment support in exchange for sovereign deployment case studies. Sovereignty guarantees (data residency, no training on your data, regulatory compliance) are locked into every agreement from the outset.
Pilot
Deploy to a meaningful cohort. Iceland targeted 10% of its K-12 teachers. Structured professional development, dedicated support, and feedback loops from day one.
Validation
Independent research partnerships measure outcomes. International peer review tests the methodology. Case studies are collected for international dissemination.
Strategy & Scale
Use validated evidence and partnership outcomes to build a national AI strategy grounded in measured impact. Transition from pilot to sustainable national program through institutional partners. Expand across sectors based on demonstrated results, not assumptions.
Engagement Model
The Sovereign Sprint: Day 1 to Day 180
The Icelandic AI Model compresses national AI adoption from years to months. This is the operating cadence we run.
Readiness Assessment
Stakeholder mapping, regulatory landscape, and target cohort identification begin immediately.
Strategy Locked
Strategic specification signed off. Vendor partnerships locked, with sovereignty guarantees built in.
Pilot Live
Pilot live with a meaningful cohort. In Iceland: frontier AI in the hands of 10% of K-12 teachers.
National Scale
National-scale pilot complete with measured outcomes. Evidence base ready for national strategy and cross-sector expansion.
Every step produces an asset your government keeps: playbooks, procurement language, governance frameworks, evaluation protocols.
Capabilities
Government AI Deployment Operating System
The Icelandic AI Model: a repeatable methodology for sovereign AI deployment. From strategic advisory to technical deployment to partner orchestration, a system that transfers, not a service that depends on consultants.
National AI Strategy
End-to-end strategy development for sovereign AI initiatives
Sovereignty Analysis
Trade-off frameworks for sovereignty vs. capability decisions
Regulatory Framework
AI governance and compliance architecture design
PPP Structuring
Public-private partnership design and negotiation
Risk Mitigation
Assessment and planning for deployment risks
Model Localization
Fine-tuning and adaptation strategies for local context
Infrastructure Planning
Sovereign compute and data residency architecture
Data Governance
Compliance architecture for GDPR and local regulations
Vendor Management
Selection criteria and relationship governance
Security Framework
Security-first deployment and monitoring protocols
Partner Connection
Connecting nations with frontier AI technology partners
Sovereignty Negotiation
Negotiating data and model sovereignty with hyperscalers
Ecosystem Building
Developing domestic AI capabilities and talent
Research Programs
University and research institution partnerships
Stakeholder Coordination
Multi-party alignment across government and industry
Five Models of AI Sovereignty
A codified methodology for evaluating and implementing sovereign AI strategies
Full Autonomy
Domestic models & infrastructure
Sovereign Infra
Foreign models, domestic compute
Governed Partnership
Strong contractual controls
Strategic Access
Preferred access, policy controls
Open Consumption
Commercial APIs
Deliverable Assets
6 transferable playbooks and frameworks included in engagement
Sovereign AI Operating System
Phases, milestones, artifacts per deployment stage
Reference Architecture
Data governance, RAG, eval, logging, security
Government Deal Kit
Procurement language, DPIA templates, risk register
Pilot Instrumentation
Outcome measurement, dashboards, eval protocol
Partner Governance
Steering committee templates, escalation pathways
Safety Case Outline
Structured safety assessment for gov't deployment
Differentiators
What Makes It Different
Vendor-partnered, not aid-dependent
AI companies invest in deployments because sovereign-scale case studies are strategically valuable to them. Governments gain access to frontier technology and training at reduced cost, without relying on traditional development funding.
Sovereignty-first architecture
Data residency, regulatory compliance, and government control are built into every partnership agreement from the outset, ensuring national interests are protected throughout deployment.
Platform-agnostic by design
Iceland's pilot ran both Anthropic Claude and Google Gemini, allowing comparative evaluation. The model adapts to whatever vendor configuration best fits a country's needs and opportunities.
Built for replication
The framework adapts to country size, institutional maturity, and language requirements. BTR delivered foundational research for Zanzibar's national AI strategy and is now engaged on Phase 2, the model's first adaptation beyond Iceland.
Recognition
International Recognition
Announced the deployment BTR architected on its official blog (November 2025) as one of the world's first national AI education pilots. Read the announcement →
Anthropic's short film from inside Iceland's national teacher deployment (August 2026). Watch →
Showcased the deployment in Google's coverage of responsible AI partnerships in Nordic classrooms (December 2025). Read the post →
Program presented at OECD convenings, including the Digital Education Outlook 2026 conference on effective uses of generative AI in education, and the Slovakia AI Skilling Summit. Contributing to the EU AI Skilling Framework. Conference →
The program's classroom practice featured in The New York Times' coverage of global AI adoption in schools (January 2026). Read the article →
Nine countries across Europe and North America have formally expressed interest in learning from the program.
On Film
What the Film Shows
The film is a portrait of how Icelandic teachers felt as frontier AI arrived in their classrooms, skepticism and enthusiasm both.
The skepticism is the point. A national deployment that produced only enthusiasm would not be a national deployment; it would be a demo. The program was built to surface exactly this range of response: roughly 10% of teachers reached through structured training, dedicated support, and feedback loops from day one, rather than a tools-only rollout. The teachers in the film are describing the conditions that design was meant to handle.
Working With BTR
Deploy the Model
BTR designed, coordinated, and delivered the Icelandic AI Model, and maintains the vendor relationships with Anthropic, Google, AWS, OpenAI, Microsoft, NVIDIA, and others that make its partnership-based structure possible.
Engagements begin with a Readiness Assessment: a scoping engagement that evaluates institutional readiness, identifies target cohorts, maps the regulatory landscape, and produces an adaptation plan with cost projections and a vendor partnership strategy.
The full framework document is available on request.
Request a Readiness Assessment