AI, society and cultural sciences

As the "Summit for Action on Artificial Intelligence" is being held in Paris on February 10 and 11, 2025, bringing together some 100 heads of state, it's not discourteous to set aside time for a round-table discussion, with nothing official or unofficial about it.
Concept de transformation numérique de fond de puce de technologie d'IA
Ai technology microchip background digital transformation concept © Rawpixels / Freepik‎

Supported by Inalco's Équipe de Recherche Texte Informatique Multinguisme (ERTIM) and the collective La Reconstruction, this round-table discussion will bring together Valérie Beaudouin, director of studies at EHESS; Jean Lassègue, director of research in philosophy at CNRS; Giuseppe Longo, director of research emeritus at CNRS, ENS Ulm; François Rastier, linguist, honorary director of research at CNRS; Mathieu Valette, professor at Inalco.

Among the themes addressed

  1. The documents generated by AI depend closely on the training datasets, their authenticity and reliability - without even mentioning their legal status, and their relationship to objective reality.
  2. How can we characterize the productions of mainstream AI, in contrast to the creations of text, images and music?
  3. Doesn't official enthusiasm induce judgmental biases? It relies on the prestige of the tech trusts that are actively involved in promoting AI, which is still not very profitable despite unprecedented investment. Doesn't embracing the promises of AI amount to "intellectual dumping" aimed at making users captive?
  4. Doesn't the prestige of a superhuman "intellectual" force lead to a loss of autonomy that can go as far as a form of voluntary servitude?
  5. Finally, are the new forms of governmentality based on AI and symbolized by the mission entrusted to Elon Musk to reduce the federal state in the USA, as by the application of the Chinese program of integral social control, compatible with democracy and the rule of law?

About the participants

  • Valérie Beaudouin is director of studies at the École des Hautes Études en Sciences sociales (EHESS). Her work focuses on the design and uses of the internet and artificial intelligence, with particular attention to writing genres in the digital space. With Julia Velkovska, she co-edited the issue of the journal Réseaux on the ethics of artificial intelligence.
     
  • Jean Lassègue is a philosopher, director of research at the CNRS, and a member of the Georg Simmel Center (EHESS). Articulating epistemology, anthropology and history, his research focuses on symbolic mediations, in particular the writing systems of languages and numbers, ranging from the exact sciences to the legal sciences.
     
  • Giuseppe Longo is a mathematician of logic and calculability and an epistemologist (DRE-CNRS, Ecole Normale Sup, Paris). For the past twenty years, his work has focused on the relationship between mathematics and the natural sciences, in particular evolutionary and organismal biology. His current project develops an epistemology of new interfaces exploring historical correlations and alternatives to the new alliance between computational formalisms and the governance of man and nature by algorithms and supposedly objective "optimality methods".
     
  • François Rastier, honorary director of research at the Centre national de la recherche scientifique, is a linguist specializing in text semantics. He worked for a decade in an Artificial Intelligence laboratory. Among his works La mesure et le grain. Sémantique de corpus, Paris, Champion, 2013.
     
  • Mathieu Valette is Professor of Language Sciences at the Institut national des langues et civilisations orientales (Inalco, Paris). His work, at the interface of text linguistics and automatic language processing, falls within the field of cultural sciences. It focuses on semantic analysis and feeds into a general reflection on the mechanization of languages and cultures. Recent publications Valette, M. 2024a. "Cognitive Warfare, Culture and the National Narrative", Cognitive Engineering, ISTE OpenScience, London, 6-12. / Valette, M. 2024b. "What Does Perspectivism Mean? An Ethical and Methodological Countercriticism", NLPerspectives 2024@LREC-COLING 2024, 111-115.

Without exclusivity, two books in particular will be up for debate:

Jean Lassègue and Giuseppe Longo, L'empire du numérique. De l'alphabet à l'AI, Paris, Puf, publication March 2025.
"Digital" has become an all-purpose word whose meaning is difficult to pin down, so numerous are the announcement effects, whether to celebrate the superhuman advances it makes or would make possible or to guard against its dangers. To avoid remaining at the mercy of this ambivalence, which fluctuates according to financial investments and prevents us from thinking, we need to start by returning to the epistemological foundations of the digital age, as this is what enables us to understand its consequences, whether theoretical or social. That's why this book sets out to clarify the notion of the digital based on the three concepts at its heart: writing, calculation and the machine. It is through these that the digital revolutionizes our capacities for expression, exploration and action, by making it possible to delegate these to machines by means of calculation. This revolution in digital writing is giving rise to a feeling of intense dispossession, fuelling all kinds of fantasies, more often fuelled by financial interests than the common good. It's time to reopen the epistemological case for the digital to help make it a tool of emancipation, not subjugation.

François Rastier, The AI killed me. Comprendre un monde post-humain, Paris, Intervalles, January 2025.
The launch of the first text and image generators in the autumn of 2022 sparked an unprecedented and ever-growing craze. Artificial intelligence has invaded the public discourse where technophobes and technophiles are pitted against each other. Without getting involved in this biased debate, which shifts the responsibility for its uses onto technology, this study highlights the anthropological break introduced by the automatic generation of texts and images. It affects both the sciences and the arts, and affects the whole of social life. But it can also affect personal life, and the author takes the liberty of denying with a smile the announcements of his death being multiplied by the most popular of AI systems.