Article content
In the newest episode of The Art of Digitalisation, host Petra Holm explores what happens when governments gain a new kind of user: AI agents. She speaks with Priit Liivak, Chief Government Technology Officer at Nortal, about how public services may need to evolve when AI systems can increasingly act on behalf of people and organisations.
For more than two decades, Estonia has built its digital state around human users, supported by digital identity, interoperable registries and secure data exchange. AI agents introduce a different challenge. If an agent can communicate with government, submit information or initiate actions for a person, governments need to know who is acting, what they are authorised to do and where responsibility lies when something goes wrong.
Liivak argues that becoming AI-ready does not mean simply adding AI to existing public services. Governments first need to understand where AI actually creates value and where established solutions such as interoperability, data exchange and traditional automation remain more appropriate. Estonia’s existing digital infrastructure provides a strong foundation, but the next step requires careful decisions about identity, authorisation, privacy and accountability.
The conversation also explores different ways AI agents could interact with government. Citizens may use their own AI assistants, governments could provide agents themselves, or public services could offer dedicated interfaces designed specifically for machine-to-machine interaction. Each approach creates different opportunities and risks, particularly when agents handle sensitive personal or financial information.
Another key question is digital sovereignty. If citizens rely on AI services operated by global technology companies to interact with public services, sensitive information may pass through third-party systems. Priit Liivak therefore stresses the continued importance of secure interoperability: AI should not become a substitute for well-designed digital public infrastructure.
Looking ahead, the episode considers whether Estonia could become a testing ground for an increasingly agent-enabled state. Priit Liivak sees considerable potential, but cautions against making AI adoption a goal in itself. The objective should remain the same as throughout Estonia’s digital transformation: solve real problems, improve government efficiency and make services easier to use.
🎧 Listen on Apple Podcasts and Spotify to learn:
- Why AI agents could become a new class of users for digital governments.
- What governments need to consider before allowing AI to act on behalf of citizens.
- Why interoperability and automation should come before AI where they already solve the problem.
- How identity, authorisation and accountability could work in an agent-enabled state.
- Why AI agents raise new questions about privacy and digital sovereignty.
- How Estonia could experiment with AI while keeping public services secure and trustworthy.
As Priit Liivak points out, nobody yet has a complete picture of where AI agents will take us. What matters now is testing, learning and having the difficult conversations early,before technology determines the rules for us.
To stay updated on how Estonia’s digital society is evolving beyond technology and towards more human-centred solutions, follow The Art of Digitalisation on Apple Podcasts and Spotify and watch for upcoming episodes.
