Why digital infrastructure will decide the future of democratic government
AI is reshaping how societies function, and governments must decide how to build the digital systems beneath it. Andres Raieste argues that embedding democratic principles in digital infrastructure can strengthen trust and create a virtuous cycle of efficiency.
By Andres Raieste, SVP, Global Head of Public Sector, Nortal
For nearly 20 years I have been collaborating with governments pursuing one objective: improving government efficiency through information technology, while navigating what these choices ultimately mean for our societies. Today, as we stand at another major technological shift – AI – we must confront a simple but profound reality: how we design national digital infrastructure will determine whether it strengthens citizen trust or silently erodes it.
History shows that transformative technologies like writing, digitalisation, and now AI, fundamentally change how information flows, how collaboration happens, how supply chains work, and ultimately how societies operate. They shift not only productivity but governance itself. So the real question is not simply how AI improves public services; it is how AI changes our society, how we govern and how much control we truly have over this process.
Here it is useful to turn to socio-technical principles like Conway’s law. The technological decisions we make reflect in the organisation and governance, and the other way around. If we adopt too much surveillance technology, we drift toward a surveillance society. If we design digital government in a heavily centralised way, we gradually erode democratic principles and consolidate power, often without even noticing. But when democratic principles are embedded in digital government, they reinforce democracy itself.
AI exposes every weakness in the system below it
This creates a unique challenge for AI. AI is inherently an untransparent technology. At the same time, it promises extraordinary efficiency gains. The task in front of us is to learn how to incorporate a fundamentally untransparent technology into a democratic digital government in a way that supports trust rather than undermines it.
To do that, we must first understand what we even mean by government efficiency, and more importantly government efficiency for whom. Efficiency is not only about reducing cost or increasing throughput. In democratic societies, the purpose of government efficiency is to maximise the use and potential of public resources and individual contributions for the good of society.
Efficiency is a virtuous cycle: state capacity to implement policies, operational efficiency in executing them, and public trust reflecting how well we are doing. If there is sufficient trust, reforms become possible; reforms increase operational efficiency; wise use of public resources generates more trust, and the cycle continues.
There are always trade-offs. Every digital government programme requires navigating short-term and long-term benefits; sovereignty versus adaptability; open versus proprietary technologies. But we argue that there is no trade-off between long-term efficiency and democratic principles. Centralisation may bring short-term speed in specific projects, but democratic principles ultimately optimise the use of resources and individual contributions and increase competitiveness, producing better long-term outcomes. Self-organisation drives innovation and resilience. Transparency drives uptake. Redundancy and contestability drive competition. Democracies generally are more efficient per capita. We must not confuse the sheer volume of large states with per-capita efficiency.
Centralise it, and you choke innovation
Estonia is a living example. Being small, we often hear the argument that “all Estonia needs is one super-database.” But that simply does not work. It would bottleneck innovation, harm long-term efficiency and erode trust. Instead, we rely on strong checks and balances that prevent the concentration of power while pushing innovation down to individual departments. This allows faster adoption of new technology. We prefer open standards and interoperability over rigid, top-down choices. We create redundancy, multiple technological options, and bottom-up innovation rather than locking ourselves into decisions that are later difficult to adapt. These design principles drive real, measurable outcomes: faster justice, better healthcare, higher state revenue and a plug-and-play environment for adopting AI.
So how do we move toward this kind of model? First, we need a strong and trusted foundation that enables us to incorporate untransparent technologies safely. Regulations and checks and balances matter, but so does digital public infrastructure that gives citizens real-time visibility into how their data is used and whether checks and balances still work. In Estonia, we already have mechanisms that allow citizens to see when their data has been accessed. It is simple, almost surprisingly simple, but this is precisely what makes safe AI use possible in practice: transparency embedded into the infrastructure.
Second, true efficiency comes from modernising the core information systems and collaboration frameworks of government domains. Whether in justice, healthcare or urban planning, the biggest impact occurs when information flows lawfully and in real time across entire chains: police, prosecutors, courts, prisons, probation. This reduces cost, shortens processes, limits corruption risk and improves outcomes. The point is not AI replacing this, but AI accelerating it, as long as the underlying systems are mature.
Third, as AI becomes widespread, we must keep the user experience at the centre. AI-enabled public services do not need to look like chatbots. They can feel intuitive and anticipatory. But the choice of interface matters. It determines whether the citizen perceives the service as helpful or alienating.
Finally, we should be honest about the impact on democracy itself. Many civil servants already use GPT models daily. Policy documents are being drafted with AI. Digitalisation feeds real-time information into policymaking. Policy cycles are speeding up. The real bottleneck is becoming the traditional democratic process. If we cannot deliberate, consult or elect at the pace at which policies can now be generated and updated, we risk drifting into futures where decisions shift to algorithms or a smaller group of technocrats. But digitalisation also creates opportunities: secure digital voting, elements of digital direct democracy or augmented representation. These are possibilities, not inevitabilities, but they require conscious design.
And this is perhaps the most important point. These decisions, the design of digital government, the architecture of information systems, the rules for AI, may look technical, but they are also political. They may appear as policy decisions, but they are technical. They blur together. In practice, these are the decisions that determine the society we will end up living in.
Government efficiency and democracy are not opposites. They coexist. They reinforce one another. But only if we design our digital infrastructure to make it so.
These themes, along with a practical maturity model and recommendations for policymakers, are explored further in our report Government Efficiency in the Age of AI: Toward Resilient and Efficient Digital Democracies. This paper is co-authored by Dr. Mihkel Solvak (Tartu University), Dr. Ott Velsberg (Estonian Ministry of Justice and Digital Affairs) and Dr. Keegan McBride (Tony Blair Institute for Global Change), with review contributions from Dr. David Ronfeldt.
Download the paper: https://nortal.com/insights/government-efficiency-in-the-age-of-ai
About Partner Content
This content is brought to you by a Global Government Forum, Knowledge Partner.
Related Posts
