A public service can be beautifully digitized and still be difficult to use. A resident may need to understand several agencies, consent to data sharing and follow instructions written for a different life situation. The interface may load quickly while the person must do the work of joining fragmented services. Northern European research and policy offer ways to ask whether digital transformation serves people in those circumstances.
This Northern European Inner Technology research map is an Institute synthesis of selected work on participatory design, digital inclusion and human-centred public services. “Inner Technology” is the Institute’s framing for human capacities and conditions of action; it is not a recognized research field shared by all cited authors. The map identifies questions and evidence sources, not an exhaustive review of every Nordic or European project. It separates empirical findings, policy commitments and design proposals so readers can see what each source can support.
Why a regional map?
Nordic and European institutions share some digital policy settings, but they are not one uniform culture or experiment. Sweden, Finland, Denmark and Norway have different administrative arrangements and populations. The European Union’s commitments apply through particular legal and political structures, while Norway participates in different ways. A regional map can show connections; it should not erase these differences.
The Nordic-Baltic Cooperation Programme for Digitalisation 2025–2030 includes an explicit goal for secure, inclusive and human-centric digital transformation. That is a policy commitment and coordination agenda. It does not prove that every service in the region already meets the goal. The research task is to ask how inclusion and human-centredness are defined, measured and changed in practice.
The European Declaration on Digital Rights and Principles similarly places people and rights at the centre of digital transformation. It is a reference framework for policy. Its language can help evaluate a project, but a declaration is not an outcome measure. A service must still show whether people can access it, understand it and contest decisions made through it.
Strand one: participation in design
Scandinavian participatory design developed around the idea that people affected by technology should help shape it. Bødker and Kyng revisit this tradition and argue that participation must face larger questions of power and social impact. Their article is a research argument grounded in the field’s history, not a randomized test of one workshop method.
Participation can be shallow. Inviting workers to comment on a nearly finished interface is different from giving them influence over which problem is solved, how work changes and which alternatives remain possible. A useful map therefore asks who entered the design process, at what stage, with what authority and what happened when their preference conflicted with an institutional goal.
Elmqvist and a large Scandinavian research team later proposed participatory AI as an approach to human-centred AI. Their 2025 paper is a research preprint and proposal. It connects older participatory-design principles to the design of algorithmic systems, but it should not be cited as proof that all proposed methods have succeeded in deployment. It gives a vocabulary for research questions and case selection.
Participation is a condition, not a photograph
A workshop can produce images of collaboration while leaving decision rights unchanged. The stronger question is whether participants can alter requirements, challenge a model’s purpose and revisit a system after deployment. If the institution retains all consequential choices, the invitation may collect opinions without sharing influence.
Bødker and Kyng’s emphasis on issues that matter can be applied to AI projects. Who decides which tasks should be automated? Can workers question whether automation is appropriate at all? Are the people affected by errors able to report them and change the system? These are not aesthetic preferences about an interface. They concern authority, work organization and accountability.
The Institute’s synthesis is that participation should be documented as a chain of decisions. A project can record who proposed a change, what evidence was offered, who accepted or rejected it and why. Such a record does not guarantee justice, but it makes the distribution of voice more inspectable than a photograph of a consultation meeting.
Strand two: digital inclusion
A 2022 Nordic Council of Ministers-funded report on digital inclusion in the Nordic-Baltic region examines definitions, monitoring and gaps. It notes that high overall digital participation can coexist with people who struggle to access or use services. The report’s regional indicators are useful for locating questions, while their averages cannot describe every person’s experience.
Inclusion includes more than a connection and a device. A person may have both but lack a language version they can use, a stable identity credential, the time to navigate a process or a way to ask for help. A form that assumes continuous access may exclude people whose connectivity is unreliable. A system that requires one reading level may create a barrier for others.
Monitoring also has a measurement problem. A count of logins may show uptake, but not whether users completed the task, understood a decision or had to seek informal help. A low use rate might reflect poor design, limited need or a trusted offline alternative. The Nordic inclusion report’s attention to definitions is therefore central, not bureaucratic detail.
Who disappears from a successful average?
A country can perform well on a digital index while a particular group faces repeated obstacles. The European Commission’s Digital Decade country reports provide indicators and policy assessments for Denmark, Finland and Sweden. These reports can support comparisons about specified measures. They do not justify the statement that one country is simply more “human-centred” than another.
Disaggregation matters. Age, disability, language, income, geographic location and administrative status may shape access in different ways. A service evaluation should not assume that a category predicts every individual’s ability. It should ask whether the service works for the people who need it, including those with multiple barriers and those who would otherwise avoid the service entirely.
Recruitment for a usability study can reproduce exclusion if only confident digital users volunteer. Researchers can include assisted sessions, accessible formats and non-users, while respecting consent. They can also study the cost shifted onto family members or frontline staff who help someone navigate a difficult system. The service’s apparent efficiency may depend on invisible work elsewhere.
Strand three: services built around lives
The Nordic DigiGov Lab explored human-centric governance and public services across Sweden, Norway and Finland. Its 2026 report asks how government might organize around people’s life situations rather than around separate agency functions. It is a project report and design-oriented synthesis. It documents questions and approaches; it does not prove that every proposed service pattern has delivered measurable benefit.
A life event can cross institutions. A person who becomes a caregiver may need health information, employment adjustments and financial guidance. If each service is designed in isolation, the resident must become the integration layer. A human-centred redesign might coordinate information and reduce repeated explanation, but it must also protect privacy and preserve the person’s control over sensitive details.
“Seamless” is not always the right goal. A pause before sharing data can help someone make an informed choice. An appeal route can slow a process for good reason. The research question is not how to remove every step; it is which steps support understanding and rights, and which merely reflect organizational fragmentation.
Governance and the limits of interface fixes
A single service team may know how to improve a page yet lack authority to change the underlying rule. Human-centred design then meets governance. The DigiGov work emphasizes collaboration and leadership alongside service interfaces. An evaluation should identify which barriers lie in wording, technical systems, law, budgets or institutional incentives.
The Institute’s synthesis is to map the path of a real task across organizational boundaries. The researcher can note where a person repeats information, waits without explanation or loses the ability to correct an error. The map can then assign each problem to the level that can change it. This is a method proposal derived from the source themes, not a validated diagnostic tool.
The test of the redesign should include frontline staff as well as residents. A system that feels simpler to the public may create unmanageable hidden work for workers. Conversely, a process made efficient for one agency may increase burden for another. Research must follow the full chain rather than declaring victory at the first screen.
Strand four: AI as a political and practical choice
Participatory AI research reminds us that an AI system changes who interprets information, who can challenge an outcome and who bears an error. The question “Can a model do this?” is narrower than “Should this task be automated under these conditions?” A public service might gain speed while losing an opportunity for a person to explain circumstances the model cannot represent.
The Scandinavian participatory-AI proposal is valuable because it makes involvement and alternatives part of the design question. Yet a preprint’s proposed principles need testing in real deployments. Researchers should report whether participants had the power to modify or reject a system and whether the resulting design changed outcomes that matter to them.
AI evaluation should include documentation of data, error patterns, appeal routes and the work required to maintain the system. A demonstration that a model answers common questions is not enough to show that it handles rare, consequential cases fairly. Human-centredness must be assessed in the whole service, not inferred from a friendly conversational interface.
What the sources have in common
The policy documents call for inclusion and rights. Participatory-design scholarship asks who has a voice in shaping technology. Digital inclusion research asks who can actually use the result. DigiGov work asks whether public services fit people’s lives. These are related but not identical strands. A project can satisfy one while failing another.
For example, a service may be co-designed with frequent users but inaccessible to people who rarely engage online. Or it may reach a wide population while giving affected workers no say over the automation that changes their jobs. A research map makes these tensions visible so one success metric does not stand for the entire human experience.
The Institute’s synthesis is a four-question reading guide: Who can participate in design? Who can access and complete the service? Who bears the burden of coordination? Who can contest or revise consequential decisions? Each question needs a different type of evidence. None can be answered by a single national digitalisation ranking.
Building a living map
A durable research map should record the source, geography, population, method, date and type of claim. A policy programme belongs in a different column from a field study or an experimental evaluation. A proposed model belongs in a different column from an observed outcome. Readers should be able to see when a claim is an Institute interpretation rather than the finding of a cited project.
The map should also be revisable. Digital services and AI systems change. A 2022 inclusion report gives a useful baseline and conceptual structure, while a 2025 country report may update selected indicators. Neither should be presented as a permanent picture. An update cadence can ask which sources have been superseded and whether previously excluded groups remain absent from the evidence.
Swedish terminology deserves explicit review before a Swedish-language version is indexed. English words such as “human-centred,” “participation” and “autonomy” do not always map cleanly onto local administrative and research usage. Translation should preserve the research meaning and identify when the Institute is introducing its own term rather than quoting an established one.
A sample research question
Imagine a municipality redesigning a digital application for support during a caregiving transition. The team might first interview residents and staff, including people who do not use the existing online route. It could map the steps across agencies and identify the points at which people repeat information or lose track of a decision.
Participants could then help define alternatives, including non-digital support. The evaluation could measure task completion, understanding, appeals, staff workload and unequal effects across groups. It should record which participant proposals changed the design and which did not. This scenario is an Institute application of the mapped literature, not a result reported by one cited study.
A stronger study would compare the new route with the previous one or phase implementation in a way that permits credible evaluation. It would ask whether the service reduced burden without taking away informed choice. A faster completion time alone would not answer that question if errors or hidden labor increased.
What the evidence does not show
The sources do not show that all Northern European digital services are inclusive, that participatory workshops always redistribute power, or that a human-centred policy statement guarantees human benefit. A regional average cannot establish access for every group. A proposal for participatory AI is not outcome evidence for every deployment.
Nor do these sources validate “Inner Technology” as an independent scientific discipline. The Institute uses that phrase to connect human capacity, agency and social conditions across research areas. The map is useful when it keeps those bridges visible as synthesis and points readers back to the methods and claims of the original work.
The resident at the service desk
A service is human-centred when the resident can accomplish a meaningful task without being forced to understand the institution’s entire architecture. That is an aspiration, not a conclusion that follows from a polished screen. It must be tested with people whose lives make the task difficult, not only with people for whom the current path already works.
Northern European work offers a research agenda: share design authority, monitor inclusion, coordinate around lives and evaluate AI in its institutional setting. The map remains open to challenge and revision. Its purpose is to turn broad commitments into questions that can be answered with evidence.
Research references
- Nordic Council of Ministers. (2024). Nordic-Baltic Cooperation Programme for Digitalisation 2025 to 2030. View source
- European Parliament, Council of the European Union, and European Commission. (2022). European Declaration on Digital Rights and Principles for the Digital Decade. View source
- Bødker, S., & Kyng, M. (2018). Participatory design that matters: Facing the big issues. ACM Transactions on Computer-Human Interaction, 25(1), Article 4. View source
- Elmqvist, N., et al. (2025). Participatory AI: A Scandinavian approach to human-centered AI. Research preprint. View source
- Consortium Digital Europe. (2022). Monitoring Digital Inclusion in the Nordic-Baltic Region. Nord 2022:007. Nordic Council of Ministers. View source
- Nordic Council of Ministers. (2026). Nordic DigiGov Lab: Human-Centric Innovation for the Next Generation of Digital Government. TemaNord 2026:528. View source
- European Commission. (2025). Digital Decade 2025: Country Reports. View source