By journalist Ain Alvela.

Artificial intelligence, or AI, has recently become an integral part of our daily lives. Initially used mainly for entertainment, AI is now increasingly making its way into the software systems that drive the work of government institutions and the business world.

The use of artificial intelligence is an inseparable part of global digitalisation, and it has become clear at the political level that these processes must be consciously guided, developed, and monitored. In Europe, a so-called Digital Decade is currently underway in the public sector to build the EU’s digital future. This initiative runs until 2030 and has now reached its halfway point.

EU member states have already compiled reports on progress made so far ­showing how the transition from an analogue society to a digital one has unfolded within the European Union. These developments are also being closely monitored by the European Commission, which assesses, among other things, the strategic digital development plans adopted or updated by member states in 2023. Countries are required to submit updated versions of their national action plans, known as Digital Decade roadmaps, to the Commission every two years, taking into account the recommendations outlined in the Digital Decade report.

Estonia at the forefront of the digital transition

Estonia holds a clear leading position within the EU when it comes to the digitalisation of public services. While some European countries are well-developed in certain service areas, the overall level of digital transformation across the EU remains modest. Estonia can take pride, for example, in having ensured full patient access to e-health records already from 2024. Estonia also outperforms the European average in other areas of digital public services, including the general digital skills of its population, which are relatively strong. Reflecting this progress, Estonia has set an ambitious goal: by 2030, it aims to become the world’s most environmentally sustainable digital state with a forward-looking digital government. However, as of now, a concrete system or roadmap for how exactly this goal will be achieved has yet to be established.

The use of artificial intelligence in the business sector is also making progress – in 2024, the adoption of AI among Estonian companies more than doubled compared to 2023, rising from 5.19% to 13.89%. Estonia also ranks among the EU leaders in cloud computing, with 52.6% of companies making use of cloud services, well above the EU average of 38.7%. At the same time, Estonia’s national report acknowledges that the pace of digitalisation in the business sector remains insufficient, particularly among small enterprises. The country also lags behind the EU average in the development of very high-capacity networks and 5G mobile connectivity. Another growing concern is the shortage of IT specialists. Broader digitalisation depends heavily on the availability of highly qualified computer and software engineers, making this talent gap a significant obstacle to future progress.

A special Eurobarometer survey conducted this year shows that 79% of Estonian citizens believe the digitalisation of public services makes their lives easier. In addition, 88% say it is important for public sector institutions to play a role in curbing the spread of fake news and disinformation online as artificial intelligence continues to advance. In other words, the more AI develops, and the more services move online, the greater the need to focus on cybersecurity. In total, Estonia has €208 million available for its digital transition from the country’s Recovery and Resilience Plan, along with an additional €373 million from various EU cohesion funds, which will help accelerate the country’s digital transformation as well.

Latvia struggles with remote areas

Latvia has set the goal of providing internet access to its entire population with speeds of at least 100 Mbit/s, along with 5G coverage in all major urban areas. This aims to establish an equal and high-quality electronic communications network across the entire country. To achieve these objectives, the government adopted a development plan for the electronic communications sector in 2021, covering the years 2021–2027. This plan governs communications policy throughout Latvia. The plan defines two main areas of focus. First, the development of digital connectivity infrastructure for electronic communications, and second, the security of electronic communications networks and services. According to the plan, Latvia’s national funding for digital projects prioritises sectors where electronic communications operators are either unwilling or unable to deploy infrastructure due to economic reasons. The report submitted to the European Commission this year notes that Latvia still lags behind the EU average in the adoption of 5G, fibre-optic networks, and very high-capacity networks.

One of the key foundational documents guiding Latvia’s digital transformation is the Digital Transition Guidelines for 2021–2027. The guidelines set the goal of providing equal, fast, and high-quality electronic communications services across the entire Latvian territory to all residents, as well as to state and local government institutions and businesses. This aims to ensure the availability of communications infrastructure essential for social and economic development. The more specific objectives include ensuring that every household has internet access with a minimum download speed of 100 Mbit/s and that 50% of major urban areas (Riga, Jelgava, Liepāja, Daugavpils) and all key land transport routes are covered by 5G networks.

Latvia’s Recovery and Resilience Plan includes two measures focused on connectivity infrastructure, with a total budget of €16.5 million, representing 4% of the plan’s digital budget. These measures address so-called last-mile connectivity in rural areas and the development of passive infrastructure along the Via Baltica 5G corridor. At the same time, statistics show a significant digital skills gap between urban and rural residents. Although 5G coverage is expanding year by year, Latvia’s current network reach still falls below the EU average. The last-mile infrastructure development measure aims to provide connectivity to 1,500 households, businesses, schools, hospitals, and other publicly accessible buildings in rural areas, with a budget of €4 million. A further €12.5 million investment is planned for building passive infrastructure to ensure 100% availability of fibre-optic permanent connections along the Latvian section of the Via Baltica corridor, guaranteeing 5G coverage. Overall, Latvia has access to approximately €850 million from various EU support funds to advance its digital connectivity.

Lithuania has the most extensive 5G network

Among the Baltic states, Lithuania leads in the development of 5G networks. The Lithuanian government promises that by 2025, the international land transport corridors Via Baltica and Rail Baltica will offer uninterrupted 5G connectivity. The 5G investment project is managed by the state-owned company Plačiajuostis Internetas. However, the rollout of very high-capacity fixed networks (VHCN), such as fibre-optic cables, in rural areas is progressing less smoothly and continues to require targeted public sector investments.

In 2021, Lithuania launched a plan to develop ultra-fast broadband. The initiative aims to accelerate the deployment of electronic communications infrastructure so that by 2021–2027, internet access with speeds of at least 100 Mbit/s reaches not only households and public institutions in major cities but also those in rural areas. When planning investments, key public sector and economic activity regions, as well as public institutions, are taken into account and connected to the digital connectivity network.

In October 2021, the Ministry of Transport and Communications, together with public sector institutions and telecommunications service providers, signed a memorandum in which all parties agreed that by 2025, internet connections with speeds of at least 100 Mbit/s will be available to at least 95% of Lithuanian households.

Since it has become clear that the set goal will not be fully achieved within the planned timeframe, the plan is to increase transparency around existing infrastructure and construction activities during the implementation of the digital connectivity measures. This will be done through a new national online geographic information system. Guidelines are also being developed to simplify the legal, technical, and administrative requirements for mobile operators and to apply a flexible licensing process, focusing on local licensing, infrastructure sharing, and other market needs within the 26 GHz frequency band.

Lithuania’s Recovery and Resilience Plan supports the digital transition with a €73 million investment aimed at advancing the deployment of very high-capacity networks, including 5G and fibre-optic infrastructure, in rural and remote areas. To implement projects for developing the digital society and connectivity, Lithuania has access to just over one billion euros of EU funding.

The current breakthrough in artificial intelligence lies in machine learning

The development and use of artificial intelligence have historically followed cyclical waves, and we are currently experiencing another period of growth. The Applied Artificial Intelligence Group at the Department of Software Science in Tallinn University of Technology (TalTech) was established relatively recently. Since 2019, the group has primarily focused on implementing machine learning and other AI methods in industrial and public sector applications. Current projects include exploring the use of AI for spatial analysis and developing machine learning and AI services.

Professor Tanel Tammet from TalTech’s Department of Software Science explains that the current surge in AI was triggered by major advances in machine learning that began at least 15 years ago. According to Tammet, the rapid progress in machine learning was largely due to the mass production of graphics cards, initially designed primarily for gaming. It turned out these graphic cards were also well-suited for running machine learning algorithms, enabling computations to be performed hundreds to thousands of times faster than before.

“That enabled faster experimentation because developing machine learning is inherently an experimental science that requires testing many methods to gain new insights,” he explains. “One area that already works quite well is image recognition – in some fields, the error rate is low enough to make it practical. I believe the greatest current practical success of AI lies in image recognition programs.”

In general, there are several types of artificial intelligence. One category of AI activities is search, while another is machine learning. Broadly speaking, there are two types of algorithms – search algorithms and machine learning algorithms. Typically, relying on just one of these results in limited capabilities. Therefore, the greater potential lies in programs that aim to combine both approaches.

At the same time, Tammet admits that at this stage there are no clear goals for artificial intelligence development – experiments are underway in many different directions. “There is very little intuition about which approaches might hold the most promise,” he notes. “Over the past decade, researchers have come to realize that we cannot accurately predict how successful any specific algorithmic system will be.”

On October 15, a conference, Connected to Stay, will be held in Brussels as part of the EU Regions Week, focusing on investment in digital connectivity in Europe’s rural and remote areas.

WHAT IS WHAT: Artificial Intelligence

• Artificial intelligence, also known as AI or machine intelligence, refers to the intelligence of machines, which is commonly called machine learning.
• Artificial intelligence is a branch of computer science that studies devices capable of perceiving their environment and performing actions that increase the likelihood of achieving a specific goal.
• In general, artificial intelligence refers to a machine’s human-like abilities such as thinking, learning, planning, and creating.
• Artificial intelligence operates based on a computer program that processes data collected through sensors and cameras to make decisions accordingly.
• Machine vision is a technology that enables a machine to perceive objects in its environment, analyze them, and shape its actions accordingly to be more effective.
• 61% of Europeans view artificial intelligence positively, yet 88% believe that this technology requires careful governance.

Sources: European Parliament, TalTech, Wikipedia