Artificial Intelligence in education

The Center LEARN develops approaches and training programs to familiarize teachers with AI-based tools and methods.

The spectacular advances in artificial intelligence (AI) in recent years have led to a significant transformation in many fields. New technologies, the massive availability of data, and progress in machine learning and natural language processing have made AI indispensable. These advancements have opened up exciting new possibilities for education but have also raised numerous concerns and misunderstandings.

On the one hand, AI offers opportunities for personalized learning, improved teaching, and the analysis of educational data, enabling targeted pedagogical interventions. However, the very real challenges associated with its use, such as algorithmic biases, data privacy, and the role of these tools within education, require guidance from experts in digital education.

Indeed, to fully harness the potential of AI in education, it is essential for teachers to understand its functioning, implications, and limitations. In this regard, the Center LEARN plays a crucial role in developing approaches and training programs to familiarize them with AI-based tools and methods.


As part of these efforts, the Center LEARN is involved in interventions with various partners. Here are the different actions taken (regularly updated):

A research team conducted a survey of teachers in private and public schools in the canton of Vaud to better understand their needs and expectations. The full reports of this survey (private schools and public high schools) are available for consultation.

Professor Francesco Mondada, Academic Director of the Center LEARN, and Grégory Liégeois also delivered a conference on the “reflective model” developed in collaboration with Sociologist Dominique Boullier (Sciences Po, Paris) and a team of professors from the University of Lausanne coordinated by Prof. Boris Beaude available here. This model provides an approach to address tools such as ChatGPT.

The EPFL magazine “Dimensions” dedicated a section to the challenges of AI in education, focusing on its implications for future learningIt includes testimonials from Pierre Dillenbourg and Francesco Mondada.

Jessica Dehler Zufferey, Executive Director of the Center LEARN, joined a panel of experts titled “What future for us and for AI” during EPFL’s Open Days to debunk myths surrounding AI in education.

Professor Tanja Käser, Head of the Machine Learning for Education Laboratory (ML4ED) and also a member of the Center LEARN, was invited to the “Der Club” program on SRF – Schweizer Radio und Fernsehen to discuss the topic “Artificial Intelligence: A new era?”. She particularly addressed the role and uses of AI in the field of education.

Frédérique Chessel Lazzarotto -

The Swiss National EdTech Testbed Program

The Swiss National EdTech Testbed Program is an initiative designed to support the digital transformation in education with appropriate technological solutions.

Launched by the Swiss EdTech Collider in partnership with the Jacobs Foundation and the Gebert Rüf Stiftung, this program provides schools with the opportunity to trial innovative digital tools and provide feedback to the startups that develop them.

The aim is twofold: support teachers in identifying and implementing technology to improve pedagogy while helping startups improving their tools with feedback stemming directly from practice.

Interested teachers and schools can register directly through the program’s website.


The Swiss National EdTech Testbed Program operates on the “Test and Learn” methodology, also known as Rapid Evaluation Cycles.

This approach emphasizes quick evaluations and iterative testing, with a focus on determining what works, for whom, and under what conditions.

The testing process for each digital tool is unique, and teachers play an active role in the evaluation. This includes an introduction to the tool, hands-on testing (which can last from 1 week to 6 weeks), and a questionnaire with both open-ended and closed-ended questions.

The program analyzes the teachers’ answers and provides startups with reports to help them understand how to improve their product. Further development of the program plans to include the creation of good practice papers for teachers.

Learning Companion

The Learning Companion is the result of a collaboration between the Center for Digital Education (CEDE) and the Teaching Support Center (CAPE) at EPFL, both members of the Center LEARN.

The Learning Companion allows students to self-assess on three levels. It invites them to reflect on their learning habits, problem-solving strategies, and their project management skills.


Based on the theoretical concepts and recommendations of the book “Learning to study” (Tormey & Hardebolle, 2017), the Learning Companion is a web-based tool developed for EPFL students in 2018 but available to all university students of Switzerland.  It uses learning data collected to automatically generate personalized recommendations for each difficulty students encounter. This system also allows teachers to have access to the (anonymous) data of their class and to identify and address any relevant difficulties.

This tool offers a series of self-assessment questionnaires that allow students to situate themselves in relation to their learning habits. For example, by answering questions such as: “Do I manage my stress well?” or “Am I revising effectively?” “.

Students receive personalized recommendations and learning resources to improve their skills in the different areas.

Solving a problem in mathematics or physics requires a systematic approach that is new for some students starting their studies at EPFL.

The Learning Companion is designed to improve students’ problem-solving skills using a learning journal that they complete after working on their exercises.

The entries in the log are then visually summarized in a dashboard. This allows students to identify the areas in which they can improve, by giving leads such as “checking the calculations” or “persisting when it is difficult”.

This system allows the teacher to have access to the (anonymous) data of his class in order to identify and address the difficulties that require additional explanations.

Next step :

The Learning Companion is currently undergoing further development aimed at adapting it to the Secondary II level by the BeLEARN team.

This extension project will notably see the addition of a multilingual version, will make it possible to examine the differences between populations (compulsory school vs. university) and to prepare students to acquire self-regulation skills in anticipation of their entry into graduate studies.



In order to strengthen and extend EPFL’s contribution to the digital transformation of education, the Center LEARN has joined the partnership between the Canton of Bern’s Department of Education and Culture and five universities to form BeLEARN: a competence center whose objective is to advance digital education through a translational approach.

Within this framework, the three Bernese institutions – the University of Bern, the Bern University of Applied Sciences and the Bern University of Teacher Education – as well as the EPFL and the Swiss Federal University for Vocational Education and Training (SFUVET) are combining their competencies with a multidisciplinary network of various educational research teams working on all levels of the Swiss educational system. Start-ups specialized in digital education participate in BeLEARN through a strategic partnership with the Swiss EdTech Collider.

BeLEARN partners collaborate on research projects based directly on specific needs identified in the field and transpose their results into the implementation of practical solutions along three main axes: the development of digital skills in education, the pedagogical use of digital resources and tools, and learning analytics (data science applied to education).

Within EPFL, various projects are conducted within the framework of BeLEARN:


  • Grafmilis: The project concerns two platforms for the analysis and remediation of learning difficulties in writing, Dynamilis and Grafos, respectively developed at PHBern and EPFL. (CHILI; contact person: Pierre Dillenbourg)
  • Teaching engineering and computing ethics with Deepfakes: The project focuses on an experimental study to evaluate pedagogical approaches to develop the ethical reasoning of engineering students (CAPE; contact person: Roland Tormey)
  • FREE – Federal Library of Open Educational Resources: The FREE project aims at designing and implementing a federal library of open educational resources for different levels and curricula in Switzerland and beyond. It is complementary to the Swiss Digital Skills Academy initiative ( (REACT; contact person: Denis Gillet)
  • Learning Companion: The Learning Companion is an online tool for students to develop effective learning strategies. The project investigates the usefulness of the tool with a pre-university audience and develops a multilingual version (CEDE; contact person: Patrick Jermann)
  • Computational problem solving in MINT mobil : The MINT mobil program of the Department of Compulsory Education of the Canton of Bern includes a robotics mission, developed at the EPFL. This project aims at developing the robotics mission and measuring the impact on students and teachers. (MOBOTS; contact person: Susanne Grabowski)
  • Labnet Booster project: The Labnet project aims to assess the needs, expectations, and potential applications of networking physical and virtual laboratories. (Contact person: Jessica Dehler Zufferey)
  • “Serious Game” for Entrepreneurial skills training: In this project we design, develop and test a module for a scenario-based “serious game” to promote entrepreneurial competencies. Main target groups are primarily vocational school students and secondarily students at universities of applied sciences. (ML4ED; contact person: Tanja Käser)

The current internal call for projects is open with a submission deadline of 31.10.2023.

Jessica Dehler Zufferey -

National Research Programme (NRP77): Assessing computational thinking skills

Computational thinking (CT) is the ability to develop a problem-solving methodology that complements computers. This project studies the extent to which Swiss students are able to do so, and develops tools to measure these skills.

Portrait / Project description

We will assess the computational thinking (CT) skills in Swiss compulsory schools in three steps. First, we will design an age-based competency model for CT and identify typical problems found in each age group. In a second step, we will develop an intelligent tutoring and assessment system that can help students solve problems while measuring their computational thinking skills. In a third phase, we will test this educational robotics approach in compulsory schools in the cantons of St. Gallen, Vaud and Ticino. A special focus will be placed on the Swiss educational robot Thymio II, which is currently in the process of being introduced in schools in these cantons.


Computational thinking is the ability to break down problems into their different parts, to develop solution strategies and to present these in an algorithmic way which can be understood and executed both by humans and computers. Computational thinking is recognized as a central element in various national and cantonal education strategies. Unfortunately, very few tools are available to measure the impact of the implemented measures.


The aim of this project is to provide policy-makers and teachers with tools and protocols for the large-scale assessment of competencies in compulsory education. The project especially aims to define a competency model, to develop an intelligent tutoring and assessment system, and to test the developed tools in the cantons of St. Gallen, Vaud and Ticino.


The results will make it possible to evaluate the developed competency model; notably to determine how it can be used in schools, and how it is accepted by students, teachers, school principals and cantonal services. The findings will also enable the evaluation of the effectiveness of the different approaches to teaching CT skills. This project therefore provides a tool to appraise the impact of educational initiatives that focus on computational thinking.

Original title

Assessing the development of computational thinking skills through an intelligent tutoring system: an exploratory study in the cantons of St Gallen, Vaud and Ticino


Project leaders

Prof. Francesco Mondada, Mobile Robotic Systems Group, EPFL

Prof. Luca Maria Gambardella, Faculty of Informatics, USI Lugano

Dr. Alberto Piatti, Dipartimento formazione e apprendimento, SUPSI Locarno

Dr. Dorit Assaf, Didaktik der Informatik und Medienbildung, Pädagogische Hochschule Fachhochschule Nordwestschweiz (FHNW), Windisch

Teaching in times of covid-19

The teams in LEARN provided support for teachers at different levels in their efforts to setup and constantly adapt their distance and blended teaching. LEARN teams worked with teachers to face the sudden switch to distance teaching during the lockdown. Part of the team worked with EPFL teachers, while others focused on teachers in public primary schools.

The teams in LEARN provided support for teachers at different levels in their efforts to setup and constantly adapt their distance and blended teaching. LEARN teams worked with teachers to face the sudden switch to distance teaching during the lockdown. Part of the team worked with EPFL teachers, while others focused on teachers in public primary schools.

Our aim is to make these resources available to all teachers across institutions and levels. Some of these resources are translated in English and in French.

Support for teachers at EPFL

As a means to continously minimize the risk of transmission of covid-19, the Center for Digital Education (CEDE) and the Teaching Support Center (CAPE) have set up documentation, workhops and a helpdesk to support EPFL teachers in mixing online and on-campus teaching modes. Their ressources are regularly updated.


Support for teachers in primary schools

Since 2018, the LEARN Center has been working with the Department of Education, Youth and Culture (DFJC) of Canton Vaud on the “Digital Education” project, in collaboration with HEP Vaud and UNIL.  The phase of distance learning offered a particular challenge to teachers: to ensure continuity of learning for students and to maintain the link with families. The LEARN team wished to support them in the definition and implementation of this new task, with the tools at their disposal while remobilising the content already seen in the training that is part of the project.

We proposed, in collaboration with HEP Vaud, online sessions starting one week after the beginning of lockdown in 2020. These short and operational sessions helped to co-construct remote teaching. The ressources shared here are in French:

  1. Défis 1 – Démarrer l’enseignement à distance
  2. Défi 2 – Organiser ma première classe virtuelle
  3. Défi 3 – Animer ma classe virtuelle
  4. Défi 4 – Interroger ma classe à distance
  5. Défi 5 – Lecture d’histoires à distance
  6. Défi 6 – Parler du covid-19 avec les enfants
  7. Défi 7 – Aborder le covid-19 en classe

LEARN collaborated with Edit Change Management to create a trilogy of picture books, called Oscar & Zoe, to adress the digital challenges related to the pandemic (e.g. fake news, screen time management). Here is the news article with links towards the free download in English, French and German.



Support team for EPFL teachers -
Ressources created for primary school teachers -

Research on Education in Times of Covid-19

The two main questions our research teams dealt with concerning covid-19 and education were: How does distance teaching during lockdown affect students and teachers? How can we collect the evidence in order to inform decision-makers?

1. Research into remote teaching during the lockdown

As a reaction to the pandemic, our research teams worked to provide evidence about the impact of covid-19 related measures on education by studying teaching at EPFL and in public schools during the lockdown.

1.1. Study on EPFL teachers adapting their teaching

A study on teacher adaptability was ongoing at the moment when covid-19 hit tertiary education in February 2020. It was quickly re-designed in order to capture teachers’ adaptability in the situation of forced change. The analysis introduces perspectives of pedagogical innovation and touches upon social components of learning which, during the lockdown, have left an impact on the meaning of professional teaching role. The recording and slides of a presentation of this study in a lunch&LEARN session are linked on the left.

1.2. Study on teachers’ experiences with remote teaching in Canton Vaud

More than 5500 teachers from the Canton Vaud have participated in the study. The analysis of the factors associated with effective distance learning reveal that there are three types of factors that play a role: contextual factors (student, teacher and school equipment;  as well as the technical and socio-emotional support available), individual factors (teacher and student autonomy, competence and motivation) and pedagogical factors (practices and tools). The complete report is available here.

2. Collecting research evidence on the impact of covid-19 on education

The pandemic has had a major impact on education, not only because of the social isolation during the lockdown, but also because of the need to organise education differently after reopening of schools. New practices have emerged. Educational researchers have quickly accompanied the emergence of these new practices with research activities. Many have conducted studies to document and investigate the effects of the coronavirus pandemic on education. These studies can make a significant contribution to understanding what has happened, but also to preparing future scenarios, based on evidence.

But how to gather all these studies in order to make them accessible and useful for the entire research community? The Swiss Conference of Cantonal Ministers of Education (EDK/CDIP), the State Secretariat for Education, Research and Innovation (SERI/SBFI/SEFRI) and the EPFL Center for Learning Sciences (LEARN) teamed up to propose a platform for the sharing of research on Covid-19 impact on education in Switzerland.

2.1. Open repository

We initiated an open online repository of studies addressing Covid-19 impact on education in Switzerland and were able to gather more than 60 studies. Please feel free to use the repository for your research and share it with colleagues.

Link to access items on the platform:
Link to add your study to the repository:

2.2. Online informal conference

In December 2020, more than 100 researchers participated in an informal conference to share the current knowledge about covid-19 in education. 17 studies were presented by colleagues from UniFR, FORS, EHL, UniZH, PHZG, FSO, SFIVET, UZH, UniGE, UniDistance and EPFL. Full programme


Studies on the impact of covid-19 on education

Digital Education (EduNum)

Flagship project of the Center LEARN, its main objective is to train students to master the three pillars of digital education (computer science, uses and media).

The Digital Education project dubbed “EduNum” is a large-scale educational reform initiative in which the Center LEARN has been mandated by the Department of Education, Youth and Culture (DFJC) of the Canton of Vaud to pilot the introduction of this new subject across the entire administrative region’s school system.

Flagship project of the Center, its main objective is to train students, digital citizens of tomorrow, to master the three pillars of digital education (computer science, uses and media), to encourage creativity, to ensure the emancipation of pupils and to enhance the potential for educational diversification.

These ambitions are supported by collaboration between the major Vaud institutions specializing in education, in particular, the HEP Vaud, the University of Lausanne, the Educational Department (DP) of the canton, and EPFL via its Center LEARN’s experts.

This project is based on a 2-year continuous training of teachers, with a pilot phase aimed at measuring and making the necessary adjustments before deployment.

More details and updates on the project’s evolution are posted here on a regular basis.

Want to find out more?


Interactive Learning environments for pharmacy assistants can encourage the students to test their knowledge in consultation situations without the fear of making mistakes. Additionally, these environments can be used to collect meaningful data to analyze student’s behaviors and provide exclusive insights to teachers.

Interactive learning environments (ILE) such as PharmaSim allow students to explore their subject matter through a realistic consultation scenario while mitigating the consequences of mistakes. PharmaSim offers a motivating learning opportunity in which students can be trained to actively engage in counselling sessions, to react (critically) to the information at their disposal and to identify cross-selling opportunities.

Pharmacy assistants typically will need to apply the learning content they have acquired at school in real life consultation situations at the pharmacy. Teaching strategies must foster the transfer of theoretical knowledge to applied competencies. While role play is one way to simulate real world scenarios, many students tend to avoid risks in group work settings. Additionally, it is difficult to engage learners in learning activities regarding subject matters they can not relate to. PharmaSim is a solution that offers a protected learning environment and enhances motivation through gamification.

Christian Giang -
Peter Buhlmann -
Want to find out more?


We develop learning opportunities around Interactive Simulations that allow students to complete experiments online and explore concepts while sitting at their desk…

The use of interactive learning environments (ILEs) to engage students in inquiry- and scenario-based learning is of great pedagogical potential. Especially when such approaches are not a viable option in the physical world (e.g. because they would be too dangerous, time-consuming, or expensive), ILEs represent an interesting alternative for students to learn through exploration and inquiry as a supplement to deductive reasoning.

HeatingSim is a simulation for heating technicians, which helps to impart the technical functionalities and the interdependencies of the different components of complex heating systems.

ChemLab allows laboratory technicians to explore concepts and theories by performing experiments online – without the need for a physical science lab. The project uses the PhET Interactive Simulations developed by the University of Colorado Boulder.

Peter Buhlmann -
Christian Giang -
Want to find out more?

MOOC – Discovering Digital Sciences with Thymio

Taking your first steps in the digital world is possible!

The MOOC titled “The Thymio robot as a tool for discovering digital sciences” is the result of a fruitful collaboration between education experts which offers a smooth introduction into the fascinating world of digital technology.

To get started, you don’t need to have a Thymio robot at hand: thanks to its virtual version, you can immediately get to the heart of the matter.

As for its physical version, it will allow you to experience all aspects of interaction and programming that this educational robot has to offer.

The educational materials (videos, texts, images, exercises, quizzes and many more resources) supplied in this MOOC have been carefully selected with the aim to provide learners with a captivating journey into the digital world. What’s more, all the educational materials can be transposed and applied within a classroom context.


The French version of this MOOC was jointly designed by EPFL and Inria and adapted into German through a cross-institutional and cross-organizational collaboration (ETHZ, PH Luzern, SUPSI, PH Bern, PH St. Gallen, PH Schwyz, amXa, Bischoff). It is available on several platforms such as,,  The Italian version is under development.

The course is divided into six chapters:

Chapter 1: Introduction to Computer Science and Robotics

Chapter 2: Thymio

Chapter 3: First steps in programming with Thymio (VPL)

Chapter 4: Programming Thymio with Scratch

Chapter 5: Programming Thymio with Aseba Studio

Chapter 6: Educational activities with Thymio

The first chapter introduces the basic concepts of computer science and robotics. The second chapter is an introduction to the Thymio robot. Chapters 3, 4 and 5 cover programming the robot in different environments with increasing complexity. Finally, Chapter 6 is a practical application of the elements presented in the MOOC.

Want to find out more?

E-Puck – The Ultimate Tool to Teach Embedded Robotics!

After more than 15 years and multiple generations, the Swiss-designed e-puck is still used worldwide by more than 4’000 universities and research centers. This modular mobile robot is a key tool to teach mobile robotics.

The e-puck is an educational robot that helps generations of students learn about embedded systems and robotics. First developed at EPFL in 2004 by Francesco Mondada and Michael Bonani, a new version was released in 2018, produced by GCtronic in Ticino.

The e-puck contains 15 sensors including 4 microphones, a color camera, 8 infrared proximity sensors, a time-of-flight sensor, an inertial measurement unit, in addition to speakers, 8 red LEDs, 4 RGB LEDs, and many more features to explore. Its simple structure with 3 contact points and two wheels makes it mobile, leading to a wide range of possible uses. All of the necessary resources are available on the GCtronic wiki for e-puck2!

Although the e-puck is mainly used for educational purposes, its completeness and modularity led it to be part of research in various other fields including collective and evolutionary robotics, but also in artistic performances!

Daniel Burnier -
Matthieu Broisin -
Want to find out more?