Publish at April 29 2026 Updated April 29 2026
So far, I've refused to take stock. Didn't 88-year-old Renzo Piano say: "What you have in life is what you still have to do"? Can you imagine Renzo Piano regretting Beaubourg?
What I'm going to try to develop here is not an assessment, but rather a questioning that the reading of Arnaud Levy's little book "L'échec du numérique éducatif" provoked in me: the story of a reversal, and of what, despite everything, is worth saving.
I'll avoid mentioning the abundant techno-critical literature and its apocalyptic discourses on the ravages of screens, as the majority of these articles or books come from authors who don't seem to know much about digital culture, and who took the ambulance chase after the fact.
Arnaud Levy (1) is not one of them. An EdTech professional, developer and practitioner of interactive design, trained in digital humanities, he has himself progressed through MOOCs and invested in active pedagogy and the skills-based approach alongside Jacques Tardif (2).
The indictment he makes in his book "L'échec du numérique éducatif" is hard to dismiss.
In recent years, it's not so much that digital education has disappeared from the French political agenda, as that the messianic rhetoric surrounding it has silently evaporated. The strategic documents of 2023 speak of sovereignty and efficiency, whereas those of 2012 spoke of revolution and the citizens of tomorrow's world.
Arnaud Lévy's method is honest: he admits to cherry-picking the studies that confirm his thesis, as digital education advocates have been doing for 30 years. Why should it be up to the skeptics to prove that digital education doesn't work? It should be up to the promoters of digital schooling to prove that it is useful and safe. 30 years of tablets, interactive whiteboards and digital workspaces deployed without any real proof of effectiveness - that's the real scandal for him!
Lévy is not the first to stumble over the question of effectiveness. As long ago as 2000, Serge Pouts-Lajus, a digital education practitioner since the 90s and an advisor to local authorities, described it as "impossible": just as there are believers and atheists, there are supporters and opponents of educational technologies, and the quarrel cannot be settled by data(3). So it's with a certain paradox that Levy describes efficiency as "the dictable name of technicist and productivist ideology", while devoting two chapters to demonstrating with data that digital technology doesn't improve learning! What the serious studies say is indeed damning.
The risks documented in the book are even more serious:
The process of forced digitization underway in the French education system is presented as "a swindle, a merry trick: digital technology sprinkled with neuroscience to privatize schools and universities". For Arnaud Levy, we need to stop imposing digital technology. Not ban it, but justify and evaluate it.
Any educator who has lived through the 2000s remembers this wave of enthusiasm. Audiovisuals, multimedia and the Internet promised a new way of writing, and opened up learning possibilities that textbooks had never imagined. Access to everything for everyone was truly a miracle for the inquisitive teacher concerned with pedagogical freedom. No more institutional shackles! School became a delightful playground.
Teachers created their own websites and blogs. The educational world was teeming with personal and collective productions. It was a blessed time to be able to edit and share at will, before the first questions arose (confidentiality, copyright, image rights). The merry Wild West was intoxicated with wide-open spaces. It wasn't long before cartographers were surveying these expanses, delineating their boundaries and organizing their fences.
Behind this effervescence lies a major omission from the very beginning: the question of use and pedagogy. The European Commission itself, in its official definition of e-learning adopted at the turn of the 2000s, took it for granted that "facilitating access to resources" was sufficient to improve the quality of learning".
For a long time, the dominant discourse was that of the digital divide: if it doesn't work, it's because society isn't yet connected enough, it's because teachers aren't trained enough, whereas young people are naturally gifted "digital natives", as Michel Serres popularized with his "Little Thumbelina".
In 2003, I asked teachers of French as a foreign language what they were doing with the Internet as part of "Cartable connecté", a portal of educational resources on the Fle.fr website. For the most part, they use it to look for resources to prepare their lessons, and use it as an enhanced library, not as a transformative pedagogical tool.
Beliefs in emancipation through tools were about to be transformed into a digital world of capture, surveillance and dependence on GAFAM, with no pedagogy, no real training, at the service of markets. Year after year, this question of the effective integration of digital technology into our practices will be raised again and again, without any conclusions being drawn.
A few years later, I started writing regularly on Thot Cursus. In January 2009, I published "Innovez, c'est un ordre!Digital pedagogical innovation had become an institutional injunction enshrined in the French teacher competency reference framework as early as 2006. In August 2010, I review the work of Bouquillon and Matthews, who demonstrate that Web 2.0 is less a cultural movement than an attempt to restore the web's image with investors after the bursting of the speculative bubble in 2003: Idéologie à la plage : sous le web 2.0, les intérêts privés.
Back in 2012, the year of Michel Serres' book and the Fourgous report (6) on language teaching, Philippe Mérieu had already seen the reversal taking place: new technologies were creating the conditions for an unhoped-for democratization of access to knowledge, but at the same time, combined with unbridled consumerism and intrusive marketing, they appeared "to be the vectors of an ever more sophisticated system for capturing minds"(7).
The formula was there, "captation des esprits", and I read it, quoted it. I knew, but I didn't want to measure the extent of the phenomenon already documented by scientific studies. Certain forces, economic as much as cultural, have practiced the aikido strategy towards techno-pedagogues: never directly confronting their convictions and dreams of emancipation, but accompanying their momentum, guiding their enthusiasm and then turning this energy against what it was carrying.
Seymour Papert (8) had nourished his constructionism from the best sources: Piaget had taught him that children construct their knowledge through action; Vygotski had shown him that this construction is fundamentally social; Bruner had reminded him that we only really learn in a meaningful context.
From this triple filiation comes a simple idea: to give everyone a tool to create something real, to program a turtle with Logo, to assemble a Lego Mindstorms robot, is to set these three springs in motion simultaneously. And this model is virtually impossible to monetize.
The detour begins with "adaptive learning", which guides each student back to his or her swimming column. It produces data, generates subscriptions and promises measurable results, reducing the learner to a profile.
The same mechanism worked on MOOCs, and this is the most destabilizing example, because I believed in them. Two models clashed: cMOOCs (c for connectivist, where knowledge resides in the connections between individuals, resources and ideas, not in the head of an expert who transmits) and xMOOCs (filmed lectures, automated MCQs, 100,000 registrants alone in front of their screen). It was this second model that Coursera and Audacity imposed in 2011-2012, much to the dismay of Georges Siemens and Stephen Downes (9 ), who had conceived the original spirit. The reason for the hijacking is simple: "connectionist" Moocs are impossible to monetize. The emergence of a community can hardly be sold by subscription. Obviously, it wasn't the most pedagogically ambitious model that won the day.
I've experienced this first-hand. The teaching teams who created MOOCs tried as much as possible to focus on activities and exchanges between learners, and put real energy into it. It was an opportunity to rethink their course, to create with colleagues, to get away from the anonymity of the lecture hall; but despite all their efforts to get students involved, they couldn't get away from the lecture logic.
This desire to transmit knowledge often exceeded their audience's attention span, and this is perhaps less a problem of training than a deep-seated resistance to identity, the transmission of disciplinary knowledge being the backbone of the teaching profession. Many teachers sincerely believe that their students should "get their money's worth", in other words, receive dense, structured, expert content. Accompanying learning, animating a community, designing situations where students construct their own knowledge is perceived as a thankless, even degrading task for those trained in the tradition of knowledge transmitted ex cathedra.
Some MOOCs were supposed to bring together adults undergoing training or retraining and "captive" students, so as to cross-fertilize and enrich each other's points of view. The graft rarely took, for two main reasons: running the forums (which was indispensable!) required human resources that were soon in short supply. Secondly, a genuine culture of peer evaluation, in the sense of an ability to give and receive constructive feedback, was lacking. This evaluation culture cannot be improvised; it has to be learned.
The conclusion is the same as what research has documented for digital technology in general: MOOCs have only benefited those who already had all the cards in their hand, the most qualified, the most autonomous, the most at ease with academic codes, those who needed the least help.
In short, it's a broader problem that digital technology has revealed but not been able to solve.
What remains is what's worth saving.
Lévy's indictment is salutary, but it sweeps too broadly. Three areas deserve to be defended:
Firstly, evaluation properly understood.
A dashboard that helps students to observe their own progress, identify sticking points and regulate their learning is formative assessment in the service of autonomy, not an instrument of control. Tools that train students to co-assess, to produce structured feedback and to construct criteria rather than be subjected to grades, are one of the most promising areas of digital education. The distinction is not between digital and non-digital, but between assessment at the service of the learner and assessment at the service of the system.
Then there's text production.
Arnaud Lévy sees the spellchecker, for example, as a tool that creates an illusion of competence in pupils and students, and short-circuits their questioning. This is to forget that teachers who underline mistakes in red without explanation or remediation often do no better! In both cases, the only way to learn is to become aware of it.
Perhaps schools don't spend enough time nurturing the skill of reflexivity, and that's where we need to put all our efforts. Denis Legros's (10) work in cognitive psychology showed as early as the 2000s that hypertext systems could be genuine cognitive aids to text comprehension and production, provided they were designed with learners' actual processes in mind, and not simply as digital libraries. This is exactly the distinction that the massive deployment of digital work environments and tablets has never made.
Real-time collaborative writing s
On Etherpad or Google Docs, where we see ourselves writing together, negotiating meaning word by word, is in my view the most beautiful pedagogical tool of the digital age. The study by Zsuzsanna Abrams (2019) in Language Learning & Technology (11) shows that truly collaborative groups produce more coherent and richer texts. It's not the technology that produces these effects: it's the peer-to-peer engagement that the technology otherwise makes possible. No textbook, no traditional classroom allows what a shared document allows in real time: to see the other hesitating, deleting, reformulating, and building together something that belongs to no one in particular.
What makes the difference in learning is the teacher. John Hattie's meta-analyses of tens of millions of students are unequivocal: among the factors with the greatest influence on success, the top ten relate directly to the teacher, to his or her relationship with learners, to his or her ability to diagnose what is really blocking.
As Arnaud Lévy says in the third part of his book: "Taking care of schools" means first and foremost training pedagogues. France recruits subject specialists and sends them into the classroom or lecture hall without having taught them much about how human beings learn. Digital technology has been massively deployed in this vacuum, and this emptiness largely explains why it has changed nothing.
The book doesn't stop at indictment, however, and the very last section sketches out the contours of a justified digital public interest, serving learners rather than markets. These militant avenues deserve careful reading, and perhaps a separate article of their own!
The same scenario is emerging for AI in education, with the same fads, the same massive investments and the same intellectual and human "laziness". The temptation to use prostheses is at its greatest, but far from being a problem of tools, it's a problem of our relationship with knowledge, which precedes digital technology by several centuries. All that remains is for us to hold this line: to work, to delve into singular situations, to resist the injunctions of enthusiasm, and not to abandon to technology that which comes under the heading of "pedagogical genius".
References
1. Levy, A. (2024). L'échec du numérique éducatif. - Le bord de l'eau - https://www.editionsbdl.com/produit/lechec-du-numerique-educatif/
2. Tardif, J. Site de l'Université de Sherbrooke. See in particular: Tardif, J. (1999). L'évaluation des compétences: documenter le parcours de développement. Chenelière Éducation.
3. Pouts-Lajus, S. (2000). Une question impossible: l'efficacité pédagogique. Available on HAL: https: //edutice.hal.science/edutice-00000101
4. Livingstone, S. (2009). Children and the Internet: Great Expectations, Challenging Realities. Polity Press. See also: www.sonialivingstone.net
5 Tricot, A. & Chesné, J.-F. (coord.) (2020, October). Numérique et apprentissages scolaires, summary report. CNESCO. Available at: cnesco.fr
6. FOURGOUS report: https: //www.vie-publique.fr/rapport/30937-reussir-lecole-numerique-rapport-de-la-mission-parlementaire-de-jean
7. Meirieu, P. (2012). La pédagogie et le numérique: des outils pour trancher? In Kambouchner, D., Meirieu, P. & Stiegler, B., L'école, le numérique et la société qui vient. Mille et une nuits. Available at: meirieu.com
8. Papert, S. (1980). Mindstorms: Children, Computers, and Powerful Ideas. Basic Books. French translation: Jaillissement de l'esprit (1981). Flammarion.
9. Siemens, G. (2005). Connectivism: A Learning Theory for the Digital Age. International Journal of Instructional Technology and Distance Learning, 2(1). See also: Siemens, G. & Downes, S. (2008). Connectivism and Connective Knowledge. Open Online Course, University of Manitoba. https://www.downes.ca/
10. Legros, D. & Crinon, J. (dir.) (2002). Psychologie des apprentissages et multimédia. Armand Colin.
11. Abrams, Z. I. (2019). Collaborative writing and text quality in Google Docs. Language Learning & Technology, 23(2), 22-42. https://doi.org/10125/44681
You might also like:
The tablet, a factor of intergenerational equity
The tablet is no longer a UFO among seniors. Its use has grown strongly among seniors for needs that are constantly evolving. Its major asset is its simplicity of use and navigation. Still, some ergonomic and application improvements are expected ... for an equitable access to digital.
The limit of "21st century skills"
The concept of competence is reaching its limits in the 21st century: capabilities, identity capital, triadic causality, three new ways of projecting oneself into present and future situations.
Being an active participant in your training in practice
Exploration of the notion of "Being an actor in one's training". Contractualization or the birth of an attitude in the long term
Trust, the currency of collaborative consumption
If there is one area in which the Internet has transformed practices, it is that of consumption. Alongside e-commerce is indeed growing an alternative mode of consumption, which uses peer-to-peer networks and is based on the capital of trust between sellers and buyers.
All the methods to become a learning territory. Little by little, each one discovers the other and understands that he or she is part of a restorative community, a community capable of building its future without depending on external subsidies.
