How to explain algorithms and train people to do so?

It has been a year since the start of our project and we are happy to share two deliverables :

• the first is a guide presenting a collection of activities and projects aimed at explaining algorithms.
• the second is a benchmark exercise which aims to advance algorithmic literacy and transparency across the European Union.

All deliverables are published through a CC BY-NC 4.0 licence.

How to explain algorithms?

A Gallery of approaches from the Algo->Lit project

This guide presents a collection of activities and projects aimed at explaining algorithms. It constitutes a resource for social/community and inclusion workers, researchers, advocates, activists or any professionals that need to present how algorithms are working in a way that is attractive for lay citizens. The range of examples could also inspire commissioners of algorithmic literacy projects such as State organisations, schools or data regulators. Note that our project is mainly targeted at communities impacted by algorithms in France, Belgium and the Netherlands.

Type of examples presented in this document:

• Collaborative workshops,
• Collective decision-making and supervision mechanisms,
• Static and pedagogical diagrams,
• Algorithm documentations and dataset stories,
• Simulators and demonstrators.

The guide can be downloaded in three languages:

• EN: Guide_Algolit_EN.pdf
• FR: Guide_Algolit_FR.pdf
• NL: Guide_Algolit_NL.pdf

Algorithmic Literacy and Transparency Competencies.

Portrait of Skills Referential

The report presents the results of a benchmarking exercise which aims to advance algorithmic literacy and transparency across the European Union. Focused on Belgium, France, and the Netherlands. At the heart of this analysis is the recognition that algorithmic literacy is a key enabler of effective transparency. Algorithmic literacy is defined not only as awareness of algorithmic presence in everyday platforms and services, but also as the ability to critically evaluate and strategically engage with these systems. Transparency, in turn, is understood as more than disclosure, it involves explainability, accountability, and the empowerment of individuals and communities to contest and reshape algorithmic decisions. The benchmarking maps and evaluates a range of twelve digital competence frameworks, trainings, and assessments, applying a structured methodology built around ten dimensions. These include governance models, purpose, accessibility, and most importantly, their contribution to algorithmic literacy and transparency, from raising awareness, to fostering critical thinking, and supporting emancipatory uses of AI.

The benchmark can be downloaded in three languages:

• EN: EN - ALGO-LIT Benchmark report 2025.pdf
• NL: NL - ALGO-LIT Benchmark report 2025.pdf
• FR: (to be accessible very soon)

For more information or questions about this work, please contact us at: algolit@datactivi.st