Estonia - AI and Data Action Plan

Overview

The AI and Data Action Plan 2024–2026 (commonly referred to as the Kratt plan) is Estonia’s operational roadmap to deliver an inclusive, human‑centred and innovation‑friendly data and AI ecosystem over a three‑year horizon. Released alongside the Data & AI White Paper 2024–2030, the plan sets out a series of coordinated measures to accelerate public sector adoption, support private sector uptake (with special focus on SMEs and deeptech), and strengthen national capabilities in language technologies and data governance. The plan was developed under the leadership of the Ministry of Economic Affairs and Communications (MKM) with contributions from sector ministries and the Government Office; its strategic framing and continuity with Estonia’s prior AI strategies are summarised on the national Kratt portal at Kratid – Vision & Strategies.

Definitions

Key terms used in the action plan reflect EU common usage and Estonian policy practice: "AI" covers systems that perform tasks with varying degrees of autonomy, including machine learning and generative models; "data governance" refers to legal, organisational and technical arrangements for responsible data use; "human‑centred" denotes approaches prioritising human oversight, rights protection and explainability; "competence centre" refers to an institutional vehicle for training, advisory services and cross‑sectoral assistance; and "high‑impact" or "high‑risk" systems are those with substantial effects on safety, health, fundamental rights or essential services. The plan intentionally aligns its terminology with contemporary EU acts and guidance to support interoperability and legal alignment.

Governance and Institutional Framework

The plan designates MKM as the primary coordination body responsible for steering implementation, monitoring progress and convening stakeholder groups, working in concert with the Government Office (Riigikantselei) and sectoral ministries. Operational support and technical implementation are assigned to public agencies such as Riigi Infosüsteemi Amet (RIA) and other specialist bodies. The plan proposes establishing or strengthening an AI competence centre (a coordinated public‑private research and support entity) to provide advisory services, testing infrastructure and piloting capacity for both public sector agencies and private actors. Governance modalities emphasize cross‑ministerial steering, public consultations with industry and academia, and an annual update cycle to keep the plan current with technological and regulatory developments.

Key Focus Areas

The Kratt action plan is structured across several interdependent pillars: (1) public sector transformation — helping government institutions adopt AI into services while safeguarding rights and processes; (2) funding and financing mechanisms — a mix of public investment, grants and support instruments aimed at catalysing development and scaling; (3) skills and education — national programmes to raise AI literacy, reskilling, and targeted support for tertiary and vocational programmes; (4) data governance and human‑centred rules — strengthening data availability, quality, consent frameworks and privacy‑preserving technologies; (5) language technology — substantial focus on resources for Estonian and other Finno‑Ugric languages to avoid digital exclusion; (6) safety, testing and certification — pilot testing, evaluation frameworks and conformity approaches in line with EU good practice; and (7) international cooperation — alignment with EU rules, participation in research networks and standardisation bodies. The plan contains concrete measures under each pillar, allocated milestones and indicative budgets for priority items.

Implementation Framework

Implementation rests on a three‑year operational structure with annual workplans, measurable indicators and designated leads for each action. MKM is charged with convening a multi‑stakeholder steering group, allocating transformation support to public bodies, and disbursing competitive funding rounds. The proposed AI competence centre is tasked with producing technical guidance, model documentation templates, a catalogue of public sector use cases and support for procurement. The plan also envisages setting up common testing environments and sandboxes for experimentation under controlled conditions. A monitoring dashboard and regular public reporting will track indicator progress and budget use.

Monitoring and Evaluation

A central monitoring framework is foreseen that combines quantitative KPIs (e.g., number of public services with AI components, SMEs supported, language resources created, training participants) with qualitative case studies and annual independent reviews. MKM will publish progress reports and update the plan each year to reflect technical developments and feedback from pilot projects. The plan emphasises transparent metrics and an iterative approach enabling course corrections. Evaluation criteria include human‑rights impact, accessibility, interoperability, cost‑benefit outcomes and measurable public value delivered through deployments.

Penalties, Liability, and Appeals

As a non‑regulatory strategy document, the action plan does not itself create new criminal or administrative penalties. It anticipates that compliance and liability are governed by existing law — notably the GDPR for data protection issues, sectoral safety and liability regimes, and forthcoming EU AI Act obligations for regulated AI systems. The plan instructs public procurers and agencies to require appropriate contractual terms (e.g., warranties, audit rights, liability clauses) when acquiring AI solutions, and to use impact assessments and model documentation as preconditions for deployment. Appeals and remedies for rights infringements remain within the established judicial and administrative channels.

Relationship to Other Instruments

The 2024–2026 plan updates and supersedes Estonia’s prior AI action documents for 2019–2021 and 2022–2023 by translating strategic goals into an operational three‑year programme. It complements the Data & AI White Paper 2024–2030 and aligns with national digital‑government strategies and sectoral policies (education, health, finance). The plan also explicitly positions Estonia to adopt and implement EU policy instruments — notably the EU AI Act, EU data strategy measures and OECD AI guidance — and to adapt national procedures (procurement, data exchange, certification) accordingly.

International Alignment

International alignment is a central pillar: the plan commits to synchronising national measures with EU legislative timelines (including preparatory actions to implement the EU AI Act), participating in European testing and certification initiatives, and contributing to standardisation efforts. It seeks to maximise Finland and Nordic cooperation in language technology and to engage with EU funding instruments to co‑finance infrastructure and research. The plan emphasises reciprocity with global norms on privacy, cybersecurity and human rights in AI deployment.

Implementation Timeline

Sources and References