Estonia is building an AI-powered state
Artificial intelligence is already reshaping economies and public services. In this interview, Kirke Maar, CEO of Eesti.ai, explains how Estonia is approaching that challenge through a new national initiative focused not on technology itself, but on how it is applied across society.
From skills and public services to business competitiveness and economic growth, she outlines how Eesti.ai aims to move beyond scattered experiments and make AI a normal part of everyday work, while ensuring the benefits are widely shared and the risks are managed from the start.
Eesti.ai is still at an early stage. What is the core problem you’re setting out to solve?
At its core, we are addressing the gap between how fast AI capabilities are developing and how slowly they are being adopted in a structured, meaningful way across organisations and institutions. In Estonia, people are already active users of AI, but this has not yet translated into workplaces, public services and core economic processes at scale.
So the issue is not access to technology, but the lack of coordinated, practical adoption that creates real value. What we are doing with Eesti.ai is moving from fragmented experimentation to systematic use that delivers measurable gains in productivity, service quality and institutional capacity.
What does “system-wide AI adoption” actually mean in practice, beyond pilots and experiments?
For us, system-wide adoption means that AI becomes part of everyday work, not something separate or experimental. It is about embedding AI into how institutions and companies actually operate.
In practice, that means public servants using AI tools in their daily tasks, companies integrating AI into their core processes, and people having the skills to apply it in their jobs. It also requires shared infrastructure, data frameworks and governance models so that solutions can scale and work together rather than remain isolated.
What have been your biggest lessons so far, even in these early stages? One of the clearest lessons is that interest is not the problem because there is a great deal of it. The real challenge is turning that interest into focused, meaningful implementation.
Another key lesson is that skills are foundational. Even where tools exist, adoption stalls if people are not confident using them.
And finally, fragmentation is a real risk. Without coordination, you end up with many small initiatives that fail to produce system-level impact. This is why maintaining a clear overall view is critical.
What role do you see for the private sector in shaping Eesti.ai from the outset?
The private sector is essential from the outset. It is not something we consult occasionally; it is a core partner in shaping and delivering the initiative.
Companies bring practical experience in deploying AI, understanding where it creates value and where it does not. They also play a central role in driving adoption across the economy and in developing skills through training and mentoring.
Through the Advisory Board and broader ecosystem collaboration, we ensure that Eesti.ai is grounded in real-world needs and not driven solely by the public sector.
Many countries are still figuring out where to even start with AI in government. What would you do first if you were advising them today?
I would suggest starting with a small number of clearly defined, high-impact use cases in government where AI can make a tangible difference. That creates momentum and practical learning very quickly.
At the same time, it is important to invest early in skills because, without that, adoption will remain limited regardless of the tools available.
Equally important is setting up a simple coordination function that maintains an overview and connects efforts across institutions. You do not need a complete strategy to begin—you need to start, learn quickly and scale what works.
How are you thinking about building trust from the beginning, before large-scale AI systems are in place?
Trust starts with clarity and responsibility. It is important to be transparent about where and how AI is being used, and equally clear about its limitations.
We place strong emphasis on data protection, privacy and accountability from the outset, and ensure that human oversight remains in place, especially in areas with higher impact on people.
Trust also builds through consistent, reliable user experience—when systems work as expected and deliver clear value.
At the same time, trust is closely linked to understanding. If people have practical knowledge of AI, they are better equipped to assess both its benefits and its risks.
What is the biggest risk right now – moving too fast, or not moving fast enough?
For Estonia, the greater risk is not moving fast enough. AI is already reshaping economies and public sectors globally, and delaying adoption would mean falling behind in productivity, competitiveness and capability.
Speed does not mean acting without control. The objective is to move at the right pace, fast enough to remain relevant, while ensuring responsible implementation and continuous learning.
What is the hardest trade-off you’re already facing – for example, between speed, control, and public trust?
The central trade-off is between speed of implementation and the need to ensure reliability and trust, especially when AI is introduced into public services that directly affect people.
There is a clear need to act and learn through implementation, but also a responsibility to manage risks carefully.
Our approach is to start with clearly defined use cases, maintain human oversight and build solutions iteratively. It is less about choosing between speed and trust, and more about ensuring they reinforce each other over time.
What is a scenario where Eesti.ai could go wrong or underdeliver, and what are you putting in place now to prevent that?
One realistic risk is that efforts remain fragmented and fail to produce meaningful system-level impact. You could have many pilots and initiatives, but without coordination, they would not lead to real change.
Another risk is that AI use remains limited to a relatively small group, rather than becoming widely adopted across workplaces and organisations.
To address this, we are focusing on a limited number of high-impact projects, ensuring strong central coordination and investing heavily in skills development. The aim is to create coherence, scale and measurable outcomes from the beginning.
Looking ahead, what would success look like over the next 2–3 years, both for Estonia and, as an example, internationally?
In the next two to three years, success would mean that AI is no longer seen as something separate, but as a normal part of how work gets done across the public sector and the economy.
We would expect to see clear improvements in productivity and service quality, alongside a measurable increase in AI adoption across organisations. Just as importantly, we would see tangible progress in skills development—at least 100,000 people participating in AI training and over one million micro-learning sessions completed, helping make AI use broadly accessible rather than limited to a small group of experts.
At the same time, this should translate into economic impact. The broader ambition is to significantly increase the value created by work in Estonia and contribute to long-term economic growth, with AI playing a central role in strengthening competitiveness.
Internationally, success would mean that Estonia is seen as a credible, working example of how to implement AI at scale in a small, open economy, focused on real outcomes, responsible use and effective public-private collaboration.
See original article here