In July 2019, an expert group led by Ministry of Economic Affairs and Communications and the Government Office presented a policy report together with proposals to advance the up-take of AI in Estonia (Estonia, 2019a). Based on the proposals, the Ministry of Economic Affairs and Communications presented Estonia’s national AI strategy (Estonia, 2019b). Estonia’s strategy provides a comprehensive overview of both existing and proposed policy measures, along with their objectives, deadlines and budget estimations. The objective of the strategy is to fully harness the potential of AI by developing and implementing policy measures in the following areas:
Encouraging the use and development of AI applications in both the public and private sector;
- Providing direct support to research in AI and increasing the relevant skills and competencies to do so;
- Developing a legal environment to facilitate the uptake of AI.
With regards to funding, the Estonian Government estimates an investment of at least EUR 10 million in 2019-2021 for the implementation of its AI strategy.
Estonia’s strategy foresees several reforms to the formal education and training systems in order to increase skills and competencies in AI. Reforms at the level of preschool, primary and secondary education will be primarily covered through an upgrade of the ProgeTiger Programme, which offers technology curriculums to schools. Awareness rising about the possibilities of AI is also supported through the technology compass offered by the Estonian Education and Youth Board. Reforms to the higher education include the Master’s programme in the field of data science and AI (in autumn 2020 the first students were admitted to the course at the University of Tartu), the promotion of elective courses on AI in postgraduate disciplines (including also non-ICT disciplines), and the increase of PhD scholarships in AI-related fields.
Additional further education trainings are in preparation and include among others online courses for citizens to raise the public awareness of AI and a training programme targeting employees of companies developing AI solutions. The Ministry of Economic Affairs and Communications has developed training courses and instructional materials in AI for managers and developers in the public sector, courses to raise awareness of AI in the public sector and data governance training courses aimed at data stewards and upper management to support the development of AI. In this respect, the University of Tartu is also providing a MOOC on AI. In addition, an Estonian language version of Elements of AI has been launched by TalTech in November 2019. The importance of providing a digital focus on lifelong learning was already anticipated in the Estonian lifelong learning strategy 2020 published in 2014.
From the lab to the market
To foster AI developments, Estonia foresees to increase the capacity of AI research. This is achieved by developing AI-related research support measures and by increasing the capacity and awareness of funding opportunities. The uptake and development of AI in the private sector will be supported through existing funding measures such as innovation vouchers, development vouchers and product development grants. Other existing policies, such as the Technology Competence Centre (TCC) and in particular the Competence Centre Specialised in Machine Learning and Data Science (STACC) are providing support measures for companies to develop innovative AI products and services.
In addition, the Estonian Government is preparing an innovation competition to promote AI developments based on governmental datasets. Another funding scheme provides financial support to pilot projects at various levels of their technology readiness. These instruments will be complemented with new funding measures to foster the digital transformation of companies (including AI and robotics) across all sectors. Flexible and sufficient funding opportunities for the AI uptake in the public sector has been earmarked through structural funds, joint procurements and new upcoming funding measures. Lastly, sandboxes are currently developed to foster testing and developing AI applications in the public sector.
To improve networking and collaboration opportunities, the Estonian Government is developing a policy tool to monitor available technology developments on the market and to liaise companies with R&D institutions. As a part of the European Digital Innovation Hubs (EDIH) network, AI & Robotics Estonia is established. The centre will be operational in June 2021 and it will allow SMEs from all sectors to obtain free services to develop knowledge-intensive solutions in the field of AI and robotics. Dissemination and uptake of AI is also targeted in the public sector among others the identification of use-cases, the organisation of meetings, conferences and a website to share experiences and good practices.
One of the actions of the AI strategy to foster public and private sector cooperation is ordering and making available AI core components, which can then be further trained by the subsequent institutions on the basis of their data and needs. The first such core component was made available through a public sector code repository in autumn 2019. A company called Texta concluded a memorandum of understanding with the Ministry of Economic Affairs and Communications to offer a free open-source text and data analysis software to all state authorities. As of 2021, there are five different core components made available to the public, including machine translation, speech synthesis, speech recognition and topic tagger. Another AI core component includes the development of an open-source chatbot established in cooperation with public bodies, spell corrector, process analyser, customer contact classifier, abstract generator, and anonymization toolkit. There are various networking events taking place to increase awareness of AI solutions, among others aimed at presenting the results of AI projects.
With respect to regulation, the Estonian strategy foresees amendments to the legislation to facilitate the development and uptake of AI. Introducing a new AI specific regulation is not planned at the moment, instead, modifications to the existing legal framework (Administrative Procedure Act) are envisaged. In addition, the Estonian Government released voluntary procurement guidelines that aim to give an overview of the most common issues as well as possible solutions that could be considered in a data science project. Finally, the Estonian Government is currently working on a self-assessment questionnaire for developers of AI that is based on the Assessment List accompanying the Ethics guidelines for trustworthy AI.
The Estonian Government is also working on national metadata standard and data quality framework, which will be implemented in a combination of guidelines and mandatory binding legislation.
The Estonian strategy devotes salient attention to data infrastructure policies. It includes data governance tools, instruments to increase the availability and responsible use of data, the creation of a Chief Data Officer in various – and potentially all – ministries, the compilation of data catalogues/sharing platforms, and the provision of funding for data audits. The project on the compilation of data catalogues has been successfully finalised and will be further developed. The data catalogue will contain all data from the public registries and will include metadata providing more information about the type of data and its availability. The open data portal will give access to all public data without restrictions. There is currently a project underway to develop a data management tool, which will initially focus on metadata management. Finally, the recommendations of the Open science in Estonia report will be further developed into cost-effective solutions for implementing open science principles at the national level.
To foster the telecommunication infrastructure, the Estonian Government continues the completion of an ultra-fast broadband fibre optic cable network and foresees activities to support the rollout of 5G network as highlighted in the Digital agenda 2020 for Estonia, released in 2018 and in the upcoming Digital Agenda 2030 (expect to be approved by the parliament in autumn 2021). In addition, Estonia joined the EuroHPC project to enjoy the benefits of supercomputing.
AI to address societal challenges
Climate and environment
In January 2020 the Estonian Prime Minister, presenting the on plans e-health, e-state and energy revolution at the World Economic Forum, mentioned the fight against climate change and the development of AI as priority pillars of the future. Following the need of finding innovative solutions to respond to global challenges such as combating climate change, the Estonian Prime Minister encouraged entrepreneurs from Estonia and the Kansai region to establish business contacts during a business seminar in Kyoto in February 2020.
Currently Estonia has several ongoing research projects that combine AI-related techniques to create a more sustainable environment:
- Ice map based on satellite date: The purpose of the ice map solution is to compile raster maps and satellite images for monitoring the snow and ice cover of Estonian coastal waters and large lakes (Lake Peipus). The maps and images are based on the European Space Agency’s as well as EUMETCast satellite data, which is made available via national satellite data centre ESTHub, in addition to in situ measurements of ice. The technology of raster maps and satellite images is based on the supervised machine learning approach, as much as possible. Ships, icebreakers, Police and Border Guard Board, Rescue Board will use ice maps in order to give warnings about ice conditions and to rescue people who have drifted with ice. Using such, already existing satellite data, reduces the need for in situ monitoring and increases the accuracy of data-based decisions. Ice map project will be finalised by the end of 2021;
- Remote monitoring of forest resources: The purpose of the project is to create an information technology solution that enables the compilation of raster maps of Estonian woody vegetation (tree species, height, forest reserves and site clearance). The maps will be constructed using multispectral satellite data, aircraft scanning data (LIDAR) and machine learning methods. The data processing should be as automated as possible and with minimal manual work by the user. The main input data for the processing are Sentinel-2 MSI multispectral satellite images obtained from the national satellite data centre ESTHub. The output of the project are four different maps: map of forest clear-cuts, map of tree species composition prediction, map of forest height prediction, map of forest stem volume prediction. With the help of such maps, it is possible to assess the forest resources and plan economic decisions on this basis. The project was finalised in July 2020;
- Monitoring and species identification using images and AI: The purpose of the project is to create a prototype based on AI (machine learning) platform for identifying animals and birds using image data (photos and videos from trail cameras). Prototype will contain machine learning models for identifying wild animals and birds in Estonia by Estonian Environment Agency. Instead of manually viewing wildlife image data (46 000 photos manually analysed in 33 hours), the AI will analyse these images in 1-2 hours. The project will begin in few months and will be finalised in 2021.
As for the COVID-19 emergency, during the first phase of the crisis, the chabot Suve helped Estonians and foreign residents of the country in finding official information about the coronavirus situation. Since then, the Estonian Government has undergone several projects and are currently developing an open-source AI-based chatbot solution. Various governmental agencies are using transcription and analysation in customer communication to understand the root causes of customer inquiries and thereby reduce the number of calls and improve satisfaction with public services.
Monitoring and future update
A working group will be set up to monitor the implementation of this action plan, to initiate additional policy initiatives if necessary and to start the preparation of Estonia’s long-term AI strategy for 2022.
Estonia (2019a). Report of Estonia’s AI Taskforce. Government Office and Ministry of Economic Affairs and Communications. https://f98cc689-5814-47ec-86b3-db505a7c3978.filesusr.com/ugd/7df26f_486454c9f32340b28206e140350159cf.pdf
Estonia (2019b). Estonia’s national artificial intelligence strategy 2019-2021. Government of the Republic of Estonia. https://f98cc689-5814-47ec-86b3-db505a7c3978.filesusr.com/ugd/7df26f_27a618cb80a648c38be427194affa2f3.pdf