As part of the recently launched Government AI Campus, Apolitical is publishing a series of articles exploring AI adoption in government. These articles take on many forms, from op-eds written by academic experts to interviews with public sector leaders working on AI adoption themselves.

Apolitical's Ula Rutkowska recently had the opportunity to speak to Dr Ott Velsberg, Estonia's Chief Data Officer, about Estonia’s innovative new AI strategy and its significant emphasis on data utilisation.

Dr Ott Velsberg is a renowned Estonian IT specialist and government official, known for his expertise in data governance and data science. As the current Chief Data Officer of Estonia, he is responsible for driving the country's data policy and initiatives related to AI and data in the public sector.

Estonia has been a leader in AI adoption. In 2019, Estonia set about implementing its Kratt strategy — named after a mythological creature — and, thereafter, devised an AI-based virtual assistant Bürokratt that helps the public access government services in a channel and device-agnostic way.

The AI landscape, however, is constantly evolving. The 2023 release of ChatGPT brought generative AI into the global spotlight. In response, Estonia has been formulating a new AI and data strategy and action plan, crafting an approach to AI adoption and governance that is designed to be flexible, adaptable and responsible.

💬 Let the author know what you think about AI adoption in government by leaving a comment below

During the interview, Ott Velsberg shared insights into the latest AI strategy and his experiences with AI implementation. What emerges prominently from this conversation is the critical role of data in any effective AI strategy. Continue reading for valuable insights and tips from Ott Velsberg on navigating the dynamic world of AI.

I understand you’ve spent the past few months working on Estonia’s new AI strategy. Can you tell me about it and what differentiates it from previous strategies?

After I joined the government in 2018, we started working on the government's first AI strategy. Since then, we have iterated on it twice. However, these iterations were basically action plans focussing on the short term — the next two or three years — due to the rapidly evolving nature of AI. Right now, we’re in the process of developing a new AI and Data strategy with a broader scope that covers the next seven years. It will define the government’s key strategic priorities and cover aspects of the private sector and society as well. It looks at the whole ecosystem of AI and data. These will still be accompanied by action plans focusing on the next three years.

Our approach is multifaceted, considering the impact of data and AI across society. We're focusing on enhancing governance efficiency, boosting private sector competitiveness and ensuring trustworthy, transparent data processing. Our strategy is built on three main pillars: a data-driven economy and society, an AI-driven government and society and trustworthy, human-centric AI and data governance.

By 2030, we envision a government where every individual has personalised digital assistants for education, mental health and other areas. Doctors, police and other professionals will have specialised assistants, optimising human work while automating simpler tasks.

The core belief is that effective data utilisation and smart data management will lead us to a top-tier data economy and robust data governance. Our three-year action plan is designed to implement this vision, with annual reviews to ensure relevance and effectiveness in improving public sector efficiency. This strategy isn't just about deploying AI; it's about creating tangible value and impact. It necessitates agility and a willingness to adapt to changing conditions and to rethink organisational approaches to AI. For the private sector, our goal is ambitious: to see 75% of companies implementing AI by 2030. A similar target is set for the public sector, aiming for AI implementation in 75% of the thousands of public sector organisations.

You've mentioned the importance of governments being smart users of data. Could you explain what effective data use entails for governments? Additionally, how does this new strategy advance beyond your previous emphasis on open data?

Open data is just one part of our data utilisation strategy. Estonia has significantly advanced in this field, as evidenced by our jump in the OECD OURdata index from 24th to 3rd place and our rise to 4th place in the European Open Data Maturity Index. We’ve learned that you need to have a holistic approach. We now have additional data exchange services that are the core of the government’s whole approach to data, we have open data and we’ve built a data scientist environment which gives more opportunity to analyse data for research purposes. We also have a consent service, which gives citizens and companies the ability to make decisions about how their data is used and shared with non-governmental organisations. Right now, we’re working on a next-generation data economy model and architecture, which among others, foresees the implementation of privacy-enhancing technologies and the development of a data marketplace and data spaces.

In Estonia's AI strategies, there's a consistent emphasis on the interplay between data and AI. Could you elaborate on why considering data is crucial in any AI strategy?

Data openness, management and privacy are fundamental components of our AI strategy and its execution. These elements are critical for the effective implementation of AI. It is also crucial to understand the impact of our actions. We need to be data-driven ourselves as well. For instance, we are currently evaluating the size of the data economy for ongoing monitoring purposes and to better plan our investments and actions. This assessment aims to better understand the contributors to the data economy, such as the expected returns from government investments in sectors like agriculture or healthcare.

As a crucial part of our strategy, every public sector organisation is required to develop its own data and AI strategy. This includes having an action plan, appointing responsible individuals at the ministerial and agency levels and establishing data management and analytics units to support each organisation. This represents a comprehensive transformation that needs to be implemented at pace. We need to have holistic operational models in place throughout the government to truly transform the way government functions.

Why is adopting an AI-first approach essential for modern governments in the 21st century?

By 2030, we envision a government where every individual has personalised digital assistants for education, mental health and other areas. Doctors, police and other professionals will have specialised assistants, optimising human work while automating simpler tasks.

Currently, government operations rely on disparate information systems and web applications. This approach is changing. We recognise the need for governments to offer services through preferred channels by individuals, reflecting the diverse ways people interact with services like banks. For example, if someone’s credit card is stolen, they call the bank instead of using online banking. Similarly, currency exchange might be done in person, while credit limit changes are made online. The government should adopt this flexible, channel-agnostic approach, allowing people to choose how, when and where they interact with government services.

Moreover, the government's approach should be proactive, personalised and centred around individual rights and preferences, especially regarding data processing and sharing. People should have the right to decide how their data is being processed, how the government provides it, how proactive the government can be in service provision and even what type of services can be provided. Data and AI are at the core of this transformation.

While there is often a focus on the negative aspects of AI, I believe there's enormous transformational potential in all sectors like healthcare and manufacturing to improve lives, outweighing the risks. However, realising this optimistic future requires solid foundations, an understanding of risks and increased societal data literacy. Every person should understand how technology affects them and their role in ensuring data accuracy and timeliness. This responsibility also brings the right to decide the extent of their data's use.

Your focus on data literacy among citizens is notable. Recently, Singapore implemented an initiative giving every citizen over 25 years old $500 to invest in skill development. Looking ahead to 2030, does Estonia have a similar plan to enhance citizens' future skills, particularly in data literacy?

The primary goal is for 80% of the Estonian population to possess foundational data literacy skills and knowledge. This means integrating data literacy from kindergarten through higher education and into lifelong learning. Data literacy should be a key component of all studies and various interdisciplinary courses, ensuring it permeates all levels of society and leaves no one behind.

From a young age, children are already engaging with technology like iPads and iPhones. It's essential to teach them, even at that early age, about the risks and opportunities associated with technology use. This education should include guidelines on appropriate interactions using technology and the importance of data privacy and security.

By building a universal skill level and knowledge base, we can draw more people into the fields of data science, analytics and data governance. Estonia currently faces a significant shortage of over 12,000 data experts, a substantial number for a country of our size and especially if you consider that from the total employment, 6.6% are already ICT specialists. This shortage is reflective of a global trend, emphasising the urgent need for action in developing data and digital skills. This skills gap is increasingly becoming a critical issue for governments worldwide.

What do you think is the most misunderstood challenge for AI adoption in government?

People sometimes think that implementing AI is a purely technical question and, in turn, take it for granted. You need to put in the hard work. Achieving long-term success in AI requires full organisational commitment, a solid structure and a clear understanding of the roles of business and data scientists. It's not merely about deploying technologies like GPT.

Practically, it's crucial to collaborate effectively with government organisations that are looking to implement AI. These government organisations have domain expertise and thus play a key role in helping to define project scope and validate results. They assist data scientists and analytics in understanding the purpose and context of data collection. However, aligning their expectations and willingness to contribute to an AI project can often be challenging.

This issue is not unique to Estonia. In my experience working with various countries, I've observed a common desire for quick wins without the necessary investment in hard work by the non-IT specialists. This pattern is a global challenge, reflecting a widespread underestimation of the complexities involved in AI implementation.

Done reading? Make sure to share your own tips by leaving a comment below ⬇️

(Image credit: Unsplash)