The Innovantage podcast is continuously expanding its horizons. While the previous episode was devoted to the experience of Lithuania in the AI revolution, this time, podcast host Sigli’s CBDO Max Golikov has invited Dr. Ott Velsberg to talk about the path chosen by Estonia.
For the last 6 years, Dr. Velsberg has been serving as the Government Chief Data Officer of the country, which allowed him to accumulate unique experience in this sphere. Ott has been overseeing such domains as data governance, open data, artificial intelligence, data privacy, and others from both strategic and practical implementation perspectives. This gives him a comprehensive vision of everything that is related to data and AI within the Estonian government. And in his conversation with Max, he shared his insights.
Check out the full Innovantage episode with Dr. Ott Velsberg here:
Estonia is one of the examples of how digital governance can be run. But it is not going to stop where it is at the moment. One of the country’s priorities for 2030 is to build an AI-powered government.
The focus is not necessarily on emerging technologies. The focus is on data itself and its value for the government, businesses, and society in general. Data plays an important role in delivering better services and making more informed decisions at different levels.
According to Ott, in the data economy, AI is one of the key pillars in boosting the growth of the economy in general.
Today Estonia already has one of the highest data economies in terms of percentage of GDP globally. It occupies the second position, just behind the United States.
However, Estonia is the leading implementor of AI in the world, as Ott highlighted, not a developer.
Of course, such services as information and communication technology as well as connectivity are known for their significant contribution to the development of AI and data economy.
But in this context, it is also crucial to mention some other domains, like manufacturing in healthcare, that are currently heavily investing in this field.
Dr. Velsberg mentioned one of the studies conducted in Estonia. They evaluated the activities and approaches of companies where at least one-third of the teams were data specialists. Nevertheless, surprisingly, no specific data-driven approaches were detected. The reason is that just the presence of particular experts doesn’t guarantee transformations.
At the same time, there are some domains, like agriculture, that do not hire people to solve any data-related tasks. Instead, they are outsourcing such services. In general, a lot of processes can be enhanced with the right application of data. If we are talking about farming, all efforts can become more efficient when farmers have enough valuable information about the characteristics of the soil and fertilizers that should be used. Outsourcing can help to address such tasks. However, with this model, agricultural businesses still do not have people who can continuously drive transformation and create the necessary environment for innovations.
Ott mentioned that today we all need to have elementary data and AI skills. The society we already live in affects literally everyone. It’s impossible to avoid its influence.
If we look at the situation with domain experts, like data analysts, data scientists, data engineers, and data stewards, we will see that Estonia currently lacks around 13,000 specialists. Specifically within the government, the country is missing one-third of data analysts.
The issue is that today it is not possible to train as many specialists as needed. However, this is not a local problem in Estonia. It is a global trend.
Speaking about a wider labor market, Ott noted that new knowledge is required not only for specific data-related roles. Even such experts as project managers need to know, for example, what the difference between a typical IT project and a data science project is, what generative AI is, what an LLM is, what machine learning is, and so on.
At the beginning of next year, the government in Estonia will be launching a data literacy campaign across the country.
There are plans to introduce various topics related to data science and analytics to young students in schools and universities.
The objective is to ensure 80% data literacy by 2030. This cannot be achieved without working with the whole society at different levels at the same time. There are always some elderly people who are not open to innovations. But the more they know about technologies, the less skepticism they will have. Talking about such topics is extremely important.
Moreover, it is vital to organize various programs. A row of AI training sessions held by the Estonian government for different social groups gained tremendous popularity. All places for the course aimed at primary school students were filled just within a day.
A good understanding of the value of technologies and data is required at the organizational level as well. Today there are a lot of companies at different stages of their maturity that work with data but they do not even understand what type of data they collect, who can access it, where it is stored, and why it is actually needed. Given this, can they efficiently manage this data and fully leverage its value? It’s highly unlikely.
The same is true about people at the individual level. Many people today have no idea what data about them is collected and how it can be further applied. Moreover, they do not even think about how they can benefit from this data. Max compared this situation with the early adoption of the internet. Users got a powerful tool but they didn’t know what exactly they could do with it.
Ott explained that today’s initiatives aimed at the development of basic IT literacy in society are part of the European goals for 2030.
Governments need to start talking about data and AI as well. In this context, it is vital to educate people about new services that will appear and that don’t even have non-digital counterparts. Here, it is also necessary to inform citizens about those groups of specialists who might become unemployed before it actually happens. It is important to mention the existing risks, as well as the factors that ensure trustworthiness and transparency.
People should have a clear vision of how they can control the use of their data.
For example, in Estonia, the government introduced the Data Tracker. Its purpose is to offer citizens access to a full overview of the operations conducted with their data. The country also has a consent service. It enables people to give the state permission to share their personal data with a certain service provider, for example, healthcare organizations.
Today, the government is also working on increasing the accessibility of services and has been heavily investing in sign language and real-time speech recognition.
Another project of the Estonian government is the development of the digital twin concept, but not in a typical sense. In this case, such digital twins can visualize different real-life situations based on available data and help the government stay proactive. For example, such solutions can be useful for foreseeing the changes in the labor market. The government is always interested in keeping the unemployment rates as low as possible. By getting insights into the possible changes, the government can organize various training courses and provide practical recommendations to those who are likely to lose their jobs in the next 6 months.
Are there any risks associated with this? Definitely yes. First of all, people may be rather confused. Not everyone is ready to get such information from the government.
Secondly, job losses are often related to the bankruptcy of companies. If the government publishes some info about a company that may go bankrupt in the near future, it can negatively affect its reputation already today.
That’s why it is crucial to think not only about how to handle the data but also how to present any information to citizens.
Trustworthiness is one of the key concerns across society. Ott shares some statistics.
Women are more likely to not trust technology itself. Many of them are less active or savvy users than men.
As for the generational differences in the Baltic states, the older generation trusts the government less, however, these people trust the private sector. With younger people, the situation is completely opposite. They don’t trust the private sector but they trust the government.
The task for the government today is to stay proactive and not to wait for a person to reach out with some questions or issues. People are likely to believe more if they understand how and what the government is actually doing.
In Estonia, there are some important initiatives that are designed to increase transparency in the AI and data sectors.
Ott also shared that at the time of recording the podcast episode, they were preparing for the launch of the algorithmic transparency standard. Its introduction presupposes that everyone who has carried out an AI project within the public sector needs to openly clarify the goals and the logic behind it, as well as other important details.
Moreover, every funded project needs to implement the Data Tracker so that all the information can be available on the government portal.
One situation last year brightly demonstrated the interest from the side of citizens in getting access to such information. Around 450,000 Estonians looked for information related to the application of their data included in the population register when there appeared some concerns regarding the ethical side of its use.
When asked about any specific approaches to building the data economy in Estonia, Ott stated that the main idea is to keep everything as simple and small as possible. Instead of large-scale projects, it’s better to start with small ones that address some specific problems.
Dr. Velsberg warned: “Don’t overanalyze. Don’t overanalyze.”
It’s necessary to listen to end customers, detect their pains, and offer solutions to them.
According to him, that’s exactly what Estonia is doing. The country’s approach can be described as problem-focused, rather than tech-focused.
There is no need to implement AI just because it is AI. It is necessary to identify the processes that can be changed and do this.
A lot of AI projects today can cost between 60,000 and 70,000 Euros and their implementation may take just around a couple of weeks. But they can save hundreds of hours of people’s time.
Countries are ready to invest millions of euros to analyze some technologies. But quite often it’s more efficient and feasible just to take action than to conduct endless research.
Of course, the introduction of a regulatory framework has some pitfalls and controversies. Should AI have separate legal entity status? Who is responsible for its decisions when it misbehaves?
But without any doubt, the legal framework for the use of AI and data is critical. It’s vital to control how data is collected and used, as well as to protect people’s rights.
Nevertheless, the Estonian government has made a decision that they won’t regulate AI nationally and will rely only on European laws.
One of the advantages of this is that the EU has a large global market and if companies that adhere to its regulations decide to enter other markets, they will have a great potential for this.
Moreover, quite often, EU-wide regulation can also push countries to the introduction of a basic standard that will be applicable to everyone across all member states. It could bring benefits not only locally but also on the international level as well.
Ott admitted that when he read the first version of the EU’s AI Act, he was extremely skeptical. The document included a lot of unrealistic requirements. The latest version has been greatly updated. It offers a lot of best practices and seems to be much closer to reality.
Ott hopes that the Act won’t undergo changes in the next few years. Time is needed to understand how the rules work and whether it is possible to improve them.
The AI Act has already downplayed or eliminated a lot of risks related to the use of AI. In some potentially dangerous situations, AI can’t be implemented at all (like social scoring which is a rather popular use case shown in many films). There are no reasons to be worried.
Governments today are interested in using AI to deliver better services to citizens and facilitate a lot of processes for them. In Estonia, a lot is done to introduce proactive and personalized services.
For example, even before a baby is born, parents can solve a lot of questions. For example, they can already give a name to their son or daughter, re-register a child to kindergarten, and get full information about the social benefits related to the birth. All this helps to save people’s time because time is the most valuable resource.
The same approaches can be applied to various aspects of people’s lives, including military service and car registration.
The government should understand the needs of people and adjust all the services to them.
Very similar rules work in the business world as well.
If you are able to provide services that your client actually needs in a personalized manner you are going to be more successful than those who provide something that people don’t actually care about.
AI and data can bring a lot of benefits at different levels of our society. But only when they are properly used with the right goals. And that’s one of the key topics discussed with experts in the Innovantage podcast. If you want to learn more about it, do not miss the next episodes.
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.png)
.png)
