Thought Leadership
Estonia is developing a framework for assigning official digital identities to AI agents, potentially allowing artificial intelligence systems to act on behalf of individuals, companies, and public institutions within clearly defined boundaries.
The proposal was introduced by the newly established advisory body Eesti.ai and has received support from Prime Minister Kristen Michal. The initiative envisions giving AI agents their own digital identification numbers, enabling structured delegation of tasks while maintaining accountability and legal clarity.
The goal is to ensure that AI systems can perform digital actions on behalf of users without requiring unrestricted access to personal or organizational accounts, data, or government services.
Limited authority, clearly defined scope
Under the current concept, each AI agent would be assigned a dedicated digital identity, making it possible to precisely define permissions and responsibilities. This approach is intended to prevent over-permissioning while enabling controlled automation of digital workflows.
Potential use cases include preparing tax declarations, generating reports, and interacting with e-government systems. Permissions would be configurable in detail, allowing organizations or individuals to restrict AI actions to specific functions such as viewing data, drafting documents, or initiating payments up to a predefined limit.
Prime Minister Kristen Michal emphasized the importance of accountability in the system:
“It must be clear who is acting, on whose behalf actions are taken, what rights are granted, and who ultimately carries responsibility.”
Kristen Michal, prime minister of Estonia
Building on Estonia’s digital leadership
Estonia views the initiative as a natural extension of its established digital infrastructure. The country has long been recognized as a leader in e-government, relying on electronic identities, digital signatures, and the national data exchange layer X-Road.
With AI agent identities, Estonia aims to further evolve this ecosystem and potentially set a global benchmark for digital governance in the age of artificial intelligence. If implemented, the country could become one of the first to formally recognize AI agents as digitally identifiable entities within public administration frameworks.
Open questions around implementation
Despite the strategic direction, key details remain unresolved. No implementation timeline or concrete legal framework has been published so far. Critical issues such as liability, security requirements, and technical architecture are still under discussion and are expected to shape the next phase of development.
AI skills push for the population
In parallel, the government is expanding its national AI literacy efforts under the initiative “The Most AI-Skilled Nation.” Between April and June, 35 workshops were held across six Estonian cities, reaching around 1,200 participants.
The program aims to scale significantly, targeting 10,000 participants by the end of the year. Additional training formats are planned, including expanded learning programs and a growing network of trainers and partner organizations.
(lb)
