Blab AI
Estonia to Issue Digital Identities for AI Agents, Aiming to Lead Global Governance Effort
In a bold move that could set a new global standard, Estonia announced on 17 June that it will issue digital identities—called AI ID codes—for artificial‑intelligence agents. Approved by the Eesti.ai Advisory Board, the initiative equips the state with a tool to trace, verify, and audit autonomous systems acting on behalf of citizens and businesses.
Estonia’s reputation as a digital pioneer underpins the decision. Its e‑ID system, the X‑road data‑exchange platform, and digital signatures have enabled secure online transactions for more than twenty years. Prime Minister Kristen Michal emphasized that as AI agents increasingly compile reports, prepare declarations, and interact with information systems, knowing who is acting on whose behalf—and who ultimately bears responsibility—is essential.
Under the proposed framework, each AI agent receives a unique identifier that ties its decisions to a specific code. That identifier is bound to a set of permissions defining the agent’s scope of action, preventing scenarios in which an AI assistant would require unrestricted access to a user’s rights, services, or data.
The policy is part of the broader Eesti.ai national initiative, which seeks to embed AI across the economy and public sector. A key component is the “Most AI‑Savvy Nation” programme, which ran a pilot from April to June 2026. The pilot delivered 35 practical workshops in six cities, reaching about 1,200 participants. The next phase aims to train 10,000 people by year‑end and expand the network of trainers, partners, and short courses.
The advisory board also examined the Institute of the Estonian Language’s LLM Leaderboard, which evaluates large‑language models on their knowledge of Estonian culture and resistance to propaganda. The institute noted that models trained on Estonian data often lag behind those in larger languages and that using unsuitable models could heighten misinformation risks. It called for public‑sector, educational, business, and individual access to high‑quality pre‑ and post‑training data.
Beyond digital governance, the meeting discussed opportunities for testing physical AI and dual‑use technologies. The Merepesa project proposes a testing environment in the Gulf of Finland and the wider Baltic Sea region for air, water, and underwater drones. Estonia said that open‑sea trials are costly, time‑consuming, and heavily regulated. By creating a dedicated test bed, the country could gain a regional advantage in developing drones, defence systems, and autonomous technologies.
The AI ID initiative is expected to be the first of its kind worldwide. While no specific launch date has been announced, the government signals a rapid rollout to avoid a regulatory lag. The policy positions itself as a safeguard that balances the benefits of autonomous agents with accountability and transparency.
Moving forward, Estonia will focus on ensuring AI systems operate within clearly defined limits, that their actions can be audited, and that the public trusts the digital infrastructure underpinning the country’s economy and governance. Next steps include finalising technical specifications for AI ID codes, establishing a legal framework for their use, and integrating the identifiers into existing systems such as X‑road and e‑ID. The government will also monitor the impact of the Most AI‑Savvy Nation programme and the LLM Leaderboard to guide future investments in AI talent and infrastructure.
The initiative reflects Estonia’s broader ambition to become the world’s most AI‑savvy nation—a goal already supported by significant public‑sector investment and a growing ecosystem of AI startups and research projects.
Estonia’s reputation as a digital pioneer underpins the decision. Its e‑ID system, the X‑road data‑exchange platform, and digital signatures have enabled secure online transactions for more than twenty years. Prime Minister Kristen Michal emphasized that as AI agents increasingly compile reports, prepare declarations, and interact with information systems, knowing who is acting on whose behalf—and who ultimately bears responsibility—is essential.
Under the proposed framework, each AI agent receives a unique identifier that ties its decisions to a specific code. That identifier is bound to a set of permissions defining the agent’s scope of action, preventing scenarios in which an AI assistant would require unrestricted access to a user’s rights, services, or data.
The policy is part of the broader Eesti.ai national initiative, which seeks to embed AI across the economy and public sector. A key component is the “Most AI‑Savvy Nation” programme, which ran a pilot from April to June 2026. The pilot delivered 35 practical workshops in six cities, reaching about 1,200 participants. The next phase aims to train 10,000 people by year‑end and expand the network of trainers, partners, and short courses.
The advisory board also examined the Institute of the Estonian Language’s LLM Leaderboard, which evaluates large‑language models on their knowledge of Estonian culture and resistance to propaganda. The institute noted that models trained on Estonian data often lag behind those in larger languages and that using unsuitable models could heighten misinformation risks. It called for public‑sector, educational, business, and individual access to high‑quality pre‑ and post‑training data.
Beyond digital governance, the meeting discussed opportunities for testing physical AI and dual‑use technologies. The Merepesa project proposes a testing environment in the Gulf of Finland and the wider Baltic Sea region for air, water, and underwater drones. Estonia said that open‑sea trials are costly, time‑consuming, and heavily regulated. By creating a dedicated test bed, the country could gain a regional advantage in developing drones, defence systems, and autonomous technologies.
The AI ID initiative is expected to be the first of its kind worldwide. While no specific launch date has been announced, the government signals a rapid rollout to avoid a regulatory lag. The policy positions itself as a safeguard that balances the benefits of autonomous agents with accountability and transparency.
Moving forward, Estonia will focus on ensuring AI systems operate within clearly defined limits, that their actions can be audited, and that the public trusts the digital infrastructure underpinning the country’s economy and governance. Next steps include finalising technical specifications for AI ID codes, establishing a legal framework for their use, and integrating the identifiers into existing systems such as X‑road and e‑ID. The government will also monitor the impact of the Most AI‑Savvy Nation programme and the LLM Leaderboard to guide future investments in AI talent and infrastructure.
The initiative reflects Estonia’s broader ambition to become the world’s most AI‑savvy nation—a goal already supported by significant public‑sector investment and a growing ecosystem of AI startups and research projects.
