• 1  https://doi.org/10.4000/dms.9724 (§ 12 and more).

2You can think about it. If you cannot find the answer, you will find it in the editorial in the latest varia issue, number 44.1 and in the annex here. We ask this question once again in this new varia issue. This issue, the last of 2024, is not made up of articles on the same theme, but brings together some various submissions.

• 2  https://doi.org/10.4000/dms.5198.
• 3  https://doi.org/10.4000/dms.5203.

5It should be noted that DMS has given a great deal of space to these issues, notably in the 2020 debate section devoted to “What the pandemic is doing to educational engineering…”2 and the contribution of Emmanuelle Villiot-Leclercq.3

• 4 https://project.inria.fr/classcode/le-projet-classcode-elements-finaux/ (accessed Dec. 12, 2024).
• 5 https://pixees.fr/classcode-v2/l-association/ (accessed Dec. 12, 2024).
• 6 https://www.didapro.org/ (accessed Dec. 12, 2024).
• 7 https://theses.fr/2023BRES0047 (accessed Dec. 12, 2024).
• 8 https://theses.fr/2023UNIP7138 (accessed Dec. 12, 2024).
• 9 https://theses.fr/2024ULILH034 (accessed Dec. 12, 2024).
• 10 https://opus.u-paris.fr/catalog/book/18 (accessed Dec. 12, 2024).

6The second article, by Jérémy Le Du, Julian Alvarez and Daniel Schmitt, describes work aimed at “identifying optimal ways of integrating video game design activities into French schools with a view to developing computational thinking.” After analyzing the existing literature, the article itself describes the implementation of an experimental research program that will be given to several dozen French schools from the start of the 2022 school year. The program, called EGPDP (Educational Game & Play Design Project), is being developed jointly by researchers from the University of Hauts-de-France and the University of Helsinki. While this description is fairly comprehensive, it does not devote much attention to the work carried out in France. For example, the Class Code project,4 for teacher training, and the collective Class’Code,5 the Didapro conferences,6 recent theses of Matthieu Branthome,7 Sevastiani Touloupaki,8 Marielle Léonard,9 the book issued from the ANR project IeCare.10

• 11 https://www.unicef.org/quality-research (accessed Dec. 12, 2024).
• 12 https://pisa.testportal.gov.ua/PISA-2022-Short-report_eng.pdf (accessed Dec. 12, 2024).

8The third contribution is a testimonial one, collected and edited by Daniel Peraya, from Yuliia Nenko, professor at the Cherkasy Institute of Fire Safety (Ukraine). It gives the floor to an expert in distance learning and training in times of crisis, an euphemism when it comes to Ukraine. Her testimony gives a well-informed account of the changes in training practices due to confinement, and the impact of the war on the school system. It is found that the combination of e-learning, stress and trauma leads to substantial learning losses for Ukrainian pupils and students.11 According to the PISA reference criteria, these reductions correspond to an approximate loss of half a year’s study in science, one year’s study in mathematics and two years’ study in reading.12

• 13 For example here: https://etsmtl.libguides.com/IAgenerative/biais (accessed Dec. 12, 2024).
• 14 See https://gagadget.com/fr/how-it-works/210106-linatteignable-sommet-de-lart-pourquoi-lintelligenc (...)
• 15 Bender Emily M., Gebru Timnit, McMillan-Major Angelina and Shmitchell Shmargaret (2021). On the Dan (...)
• 16 https://www.usine-digitale.fr/data-centers.N2219114 (accessed Dec. 12, 2024).

11In fact, the limits are amply described13 and biases, errors (like the six-fingered gloves14 or false references) are intrinsically linked to the very functioning of this technology which cannot understand the texts it manipulates. All this is well summed up in the term stochastic parrot,15 metaphor illustrating the use of statistics (probabilistic models) to convincingly generate texts similar to those of humans, but without understanding their meaning. We would like these AIs to be able to explain what they produce, which they cannot do properly. Explainability is a fundamental issue. We must not forget the cost of these technologies and the energy resources needed to run them,16 quality of service will be heavily dependent on what users are able to pay.

• 17 https://www.nytimes.com/multiplying-at-a-rapid-pace.html (accessed Dec. 12, 2024).
• 18 Readers can refer to various articles that have appeared in DMS, demonstrating the scepticism of th (...)

14The last great revolution announced, but only in the field of education, was the mooc revolution. They made the news in 2012.17 With a very large number of participants, the collection and analysis of interaction data from these participants should enable continuous improvement of the courses, making it possible to generate the best possible courses in just a few years. This culminated in a promise in the form of an oxymoron: the massive personalized course. In 2024, the horizon seems to have shifted!18

• 19 Bruillard Éric (2014). Les utilisateurs des MOOC : quel regard ? DMS 7 2014, https://doi.org/10.400 (...)

15However, the mooc had been able to restore the role of the teacher and the lecture in a sort of amphitheater of immense dimensions, since it could accommodate a very large number of people at a distance,19 sometimes with the illusion of a personal amphitheater. It is possible to establish a link between this illusory experience and the idea of “softened magistrality” (and also “getting away from a vertical word of authority experienced as arrogant”) taken up by Olivier Aïm in his interview with Claire Peltier. (Magistralité et remédiation, une rencontre « entre SIC et SED »), last article in this issue 48.

• 20 https://doi.org/10.4000/dms.9952.

16This interview concludes the series of articles devoted to the debates initiated by Claire Peltier for the “Debates” section of issue 45 of DMS,20 in examining the forms of lecturing today and approaching the lecture as a “model of transmission and relationship destined [perhaps?] to redefine itself.”

• 21 https://doi.org/10.4000/dms.9176.

17To bring this editorial to a close, we can return to the idea of trust, the solution to the riddle presented at the beginning of the text and encountered in the articles in this issue: trust in institutions, trust in machines and AI, trust in others, and so on. We should also add Chantal Bois’s earlier article on the educational alliance,21 which places great emphasis on the notion of trust (particularly the climate of trust). The notion of authority also appears. By adding the notion of legitimacy, which is central to any understanding of teachers’ activities, Pierre Rosanvallon’s latest book, “Les Institutions invisibles” (Invisible Institutions) can undoubtedly help us to better frame these three central notions for understanding education and training: trust, authority and legitimacy.