Dutch AI-in-Education Atlas
Accountability
The atlas is an interactive bubble map that shows what the landscape for AI in education in the Netherlands looks like. It is designed for teachers, educators, and school leaders who want to gain a clear understanding of this rapidly evolving field – independently, transparently, and without their personal data being collected.
Background And Purpose
The atlas is an offline, informational visualisation, not an AI system. All data are processed locally: there are no cookies, trackers, or connections to external servers. The dataset is updated manually each month, and all sources are open and verifiable. Its aim is to encourage insight and reflection, not to advise or rank. In doing so, the atlas helps education professionals better understand how AI is being applied in Dutch education, and where opportunities, risks, and collaborations are emerging.
How The Atlas Works
Each bubble in the atlas represents an organisation or initiative within the Dutch AI-in-education ecosystem. Five visual dimensions create the overall view:
- X-axis: balance between policy and classroom practice
- Y-axis: focus from technological to human-centred learning
- Size: level of impact (1–5)
- Colour: sector (policy, research, edtech, training, or bridge organisation)
- Transparency: degree of openness and independence
Users can filter by audience, sector, vendor relation, or theme, adjust contrast, or export the visualisation. Clicking on a bubble opens an information panel with key facts, strengths, limitations, and a source link. The atlas is fully accessible (WCAG 2.1 AA) and runs entirely locally in the browser.
Impact scale:
1 – local initiative · 2 – specialised domain · 3 – medium reach · 4 – broad network or programme impact · 5 – national or policy-shaping
Eight Values For Responsible AI Use
Although the atlas itself is not an AI system, its design follows the eight values for responsible AI in education.
- Human-Centred in Purpose
The tool places human needs first by helping education professionals, policymakers, and researchers understand the Dutch AI-in-education landscape without relying on commercial advisers or vendors. It does not dictate choices but informs users so they can make their own, in line with educational and ethical values. - Fair in Operation
Every organisation is presented according to the same criteria, without preference or hidden assumptions. The atlas works on all devices, is accessible to all users, and makes visible where interests or power dynamics may play a role. No automated recommendations are made – users decide what is relevant to their own context. - Transparent in Explanation
All data, calculations, and criteria are open and explainable. The five visual dimensions show how organisations position themselves, without assigning judgement. Each bubble links to a verifiable source, allowing users to check where the information comes from. To ensure full openness, it is also disclosed that the developer, Symbio6, appears as an actor within the ecosystem. The organisation is assessed using the same objective criteria, and all assumptions and data sources are public and verifiable. - Safe in Use
The atlas operates fully offline and follows the principle of privacy by design. There are no cookies, trackers, or external connections; user preferences are stored only locally. Because no personal data are processed, the tool complies with the GDPR and can be safely used in educational environments. - Reliable in Results
The data are up to date, checked, and include full source references. The atlas presents both the strengths and limitations of organisations, giving users a realistic view. Errors are corrected in monthly updates, and the dataset is openly available for review or reuse. - Accountable in Oversight
Users always remain in control. They decide which sectors, tags, audiences, or vendor types to view. The atlas performs no automatic actions or updates without user knowledge. All source code, datasets, and mappings are publicly available, allowing external review and peer validation. - Privacy-Conscious in Data Handling
The atlas collects or shares no personal information. All data remain on the user’s own device and can be easily deleted. No user profiles are created, and individuals retain full control over their information and settings. - Sustainable in Societal Impact
The atlas is lightweight, energy-efficient, and built to last. It promotes AI literacy, critical thinking, and collaboration across educational sectors. Through open data and accessible design, it supports equal opportunities and responsible technology use in education.
About The Data
The dataset contains only public information – organisation name, sector, impact level, indices, vendor relation, themes, and source link. All data come from open, verifiable sources and are stored entirely locally. The dataset is updated monthly and includes a version number and changelogue.
- Inclusion: organisations active in the Netherlands that use AI in learning or education, publicly findable and verifiable.
- Exclusion: non-educational, foreign, or inactive initiatives.
The atlas aims for representativeness, not completeness. Users can submit additions or corrections, which are reviewed and published in the following update.
Additional Best Practices
Beyond the eight values, the atlas follows additional guidelines for responsible design. It meets WCAG 2.1 AA accessibility standards: the interface can be used via keyboard or screen reader, includes high-contrast options and visible focus indicators, and uses clear, consistent navigation.
The atlas strengthens AI literacy and critical reflection by helping users see which organisations are active, how they connect, and what interests or limitations are involved. Following principles of open science and academic integrity, all sources are traceable, the methodology is reproducible, and the code is openly available. The data adhere to FAIR principles (Findable, Accessible, Interoperable, Reusable).
A do-no-harm approach ensures that the atlas presents factual, non-misleading information. Organisations are described objectively, without reputational bias or subjective judgement. Finally, the atlas complies with EU and Dutch legislation, including the EU AI Act and the GDPR. It is purely informative, non-decisional, and respects the privacy and autonomy of all users.
Qualification Under the EU AI Act
Under the EU AI Act, the atlas does not qualify as an “AI system”: it uses no machine learning, prediction, or automated decision-making. It therefore falls within the “no risk” category (informational tool). Nevertheless, the atlas aligns with the spirit of the Act by ensuring transparency, privacy protection, human oversight, and secure, traceable design. It serves as an example of how non-AI tools can still reflect AI-ethical principles.
Reflection And Inspiration
The Dutch AI-in-Education Atlas demonstrates that accountability for digital tools need not be complex. With open data, local processing, and explicit design values, even small educational projects can set an example. Developers of digital tools can learn from this approach: make decisions transparent, document sources and assumptions, disclose potential interests, and always respect the privacy and autonomy of users. In doing so, responsible AI use becomes not just a principle, but a practical design standard for everyone shaping learning with technology.