Estonia launches a digital identity project for AI agents: AI-isikukood
Estonia is taking another step towards integrating artificial intelligence into public administration. The country's government, led by Prime Minister Kristen Michal, has approved an initiative to create an official digital identity for AI agents — the so-called AI-isikukood. This project, supported by the Eesti.ai council, aims to legitimize the actions of autonomous systems in the digital space.
Why do agents need a digital ID?
The main problem that AI-isikukood solves is excessive access. Today, when delegating tasks to an AI assistant, users or companies are often forced to grant it overly broad permissions. Estonian authorities propose replacing this with limited, controlled, and auditable authorizations. The agent will be able to act strictly within the scope of its assigned rights: viewing data, preparing reports, processing payments, or operating within a set limit. According to Michal, in the future, AI will increasingly perform digital operations on behalf of humans, and it is critically important to clearly understand who is acting, on whose behalf, and who bears responsibility.
Infrastructure and timeline
The AI-isikukood project relies on Estonia's existing digital infrastructure, including the electronic ID for citizens and the e-Residency program for foreigners. A key element will be the Bürokratt platform, which from 2026 will allow each agency or domain to use its own personalized AI agent within a single cooperative network. In parallel, the Aruait project is being developed — a sovereign management layer for AI agents in the public sector, which will define the technical architecture and cooperation models. The exact launch dates for AI-isikukood, the technical architecture, and the mechanism of liability for potential losses have not yet been disclosed.
Context and prospects
The Eesti.ai initiative, launched on January 27, has already approved 15 projects with high expected impact in the fields of education, healthcare, entrepreneurship, and security. By June, some of them had moved into the analysis and market consultation phase. Estonia aims to become the first country in the world to create an official digital identity for AI agents, continuing its trajectory of digital transformation. At the same time, as cybersecurity experts, including the Anthropic team, rightly note, the safe deployment of such agents requires applying Zero Trust principles to minimize the risks of autonomous systems.
Analyst's opinion: Estonia's initiative is not just a technological experiment but a well-thought-out step towards creating a legal and infrastructural foundation for the agent economy. If the project is implemented, it could become a benchmark for other countries seeking a balance between innovation and control. However, the key challenge remains developing a transparent mechanism of liability for AI actions, especially in the financial and legal spheres.