Launched in January, Eesti.AI includes international experts, tech companies, and investors. The initiative does not replace formal government decision-making but provides strategic input to support project preparation and delivery.
Because AI will affect society beyond government and industry, Eesti.AI is also being shaped through public discussion and practical engagement. Kirke Maar, who leads the initiative, told Eesti Elu that “the goal is to make AI understandable and usable beyond a small group of experts. This means involving educators in questions of learning and skills, journalists and civil society in questions of public debate and trust, artists and creative professionals in questions of authorship and creativity, and ordinary citizens through training, feedback, and practical participation.”
Eesti.AI’s goal is to “double the value of work in Estonia by 2035, and to grow the economy by 25% within five years and by 50% within ten years.”
Why AI?
The argument behind Eesti.AI is that AI can help Estonia do more with fewer workers. “By automating routine tasks and augmenting human decision-making, AI could allow fewer workers to produce more value–in effect compressing years of productivity growth into a shorter period,” according to Estonian World.
Maar said selected projects will aim to “reduce routine work, improve service quality, increase productivity and be scaled if they work.”
In Estonia, many people have already encountered AI, yet most institutions and businesses have not integrated it systematically. Maar told ERR, “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,” she added.
According to its website, Eesti.AI’s goal is to “double the value of work in Estonia by 2035, and to grow the economy by 25% within five years and by 50% within ten years.”
Upcoming Projects
In April, Eesti.AI endorsed fifteen projects across six focus areas: people, businesses, the public sector, AI infrastructure, education, and health. Maar said these include “large-scale AI skills development for the population; AI mentoring, hackathons, vouchers, and development grants for companies; AI tools to improve law-making, public procurement, and public sector work; mapping the need for AI computing infrastructure; AI in general education and an Education Data X-Road; improving Estonian language and culture in AI models; and several healthcare projects, including AI-supported documentation in primary care and ambulance services, chronic disease prevention, and a learning health system data loop.”
The reliance on AI also raises questions about technological dependence. Some of the world’s most prominent AI companies are based in the United States, creating concern that public adoption of these tools could deepen dependence on foreign providers, especially in critical areas such as security, infrastructure and health data.
To address this, Maar said the initiative follows a technology-neutral approach. “The aim is not to favour one specific provider, model or platform, but to choose solutions that meet Estonia’s practical needs and comply with security, data protection and reliability requirements.” In some cases, she said, “the best solution may be an international commercial tool; in others, a European or open-source solution, local development, or a combination of these.”
Several projects are also focused on strengthening Estonia’s own capabilities. Maar pointed to work on Estonian language representation in large language models, stronger data frameworks in health and education, and public sector AI development as ways to reduce the risk of foreign dependence.
Another core priority is improving AI literacy. Eesti.AI plans to support practical workshops that teach people how to use AI safely and responsibly in the workplace, assess the reliability of AI outputs and avoid common risks.
“Most of these projects are planned as eighteen-month implementation efforts running until the end of 2027, with first interim results expected already by December 2026,” Maar said.
“Trust starts with clarity, transparency, and responsibility,” Maar said. “We need to be clear about where AI is being used, what it is meant to do and what its limitations are. AI should not be presented as magic or as a replacement for human responsibility.”
(Kirke Maar)
Risks and safeguards
Widespread AI adoption in critical areas such as healthcare and public services also carries significant risks. According to Maar, these include “biased or discriminatory outcomes, insufficiently transparent automated recommendations, excessive or unsafe use of personal data, erroneous AI-generated conclusions, and situations where people do not understand whether or how an automated tool has been used in relation to them.”
Eesti.AI’s broader approach is to test, learn and deploy projects responsibly. That means identifying risks early, adjusting solutions as needed, and scaling projects only when they demonstrate practical value.
Legal oversight is also part of the model. Eesti.AI introduces independent legal oversight by bringing additional capacity to the Chancellor of Justice’s Office to address AI-related legal and oversight questions.
Public trust is another central priority. “Trust starts with clarity, transparency, and responsibility,” Maar said. “We need to be clear about where AI is being used, what it is meant to do and what its limitations are. AI should not be presented as magic or as a replacement for human responsibility.”
Maar said trust will also depend on whether people see clear benefits in practice. “If people see that AI reduces unnecessary bureaucracy, improves services, saves time for doctors or teachers, and helps them in their own work, trust becomes grounded in experience rather than abstract promises.”
She pointed to healthcare, public procurement and law-making as examples. “In healthcare, AI-supported documentation in primary care and ambulance services is not meant to replace medical judgement, but to reduce manual documentation and give professionals more time for patients. In public procurement and law-making, AI is intended to reduce repetitive manual work and improve quality, while responsibility remains with people and institutions.”
~~~
This article was written by Natalie Jenkins as part of the Local Journalism Initiative.
