Gary Marcus et Ernest Davis sont deux spécialistes de l’intelligence artificielle (IA) : le premier est professeur en psychologie et neuroscience, le second en sciences informatiques. Ce livre donne un éclairage sur la situation exacte de l’IA aujourd’hui : ce que fait, mais surtout ne fait pas cette technologie. Il permet d’avoir un regard critique sur le battage médiatique auquel on assiste à ce sujet, et de mieux comprendre comment remettre l’IA sur le bon chemin pour ...
(971 more words)
With a three-part series of articles (in issues 428, 431 and 433), Futuribles journal devoted a large amount of space to understanding the human brain in 2019. Underlying many of the pieces was the question whether machines will one day outstrip human beings in their capacity to reason. But this is far from being a new question and revisiting the thinking of two pioneers in the field of human brain/machine comparison—Alan Turing and John von Neumann—offers a different view on the first machines to be inspired by, or to resemble, the functioning of the brain.
Pierre Papon demonstrates the contribution made by the two mathematicians in this area. And if the human-intelligence-versus-AI debate continues to rage, as increasingly rapid and scary technological developments come on the scene, current studies—like older ones—seem to indicate that the human brain still maintains a substantial lead over the capabilities of machines. Any number of analogies are invoked, but for the moment the complexity of the human brain seems unparalleled and unrivalled.
Selon une enquête récente, les trois quarts des salariés français sont optimistes quant à leur avenir professionnel . Cet optimisme varie cependant selon le sexe et le secteur d’activité, ainsi probablement que selon le niveau d’études bien que, de manière surprenante, ce critère n’apparaisse pas dans l’enquête en question. En revanche, celle-ci indique que 45 % d’entre eux déclarent songer à quitter leur emploi dans les cinq prochaines années (dont 22 % dans les deux ans). Ce ...
(encore 826 mots)
This third strand in the “Brain” series initiated by Futuribles in January 2019 is given over to human-machine interactions (screens, computers etc.) and, more particularly, to the impact of screens on brain function and development, especially among young people. In this article, Grégoire Borst, a specialist in neurosciences and the brain, looks into the consequences of the increased exposure to screens that has been seen in recent years, particularly among children and adolescents. While everyone is affected by this increased exposure, the question of its impact probably arises more sharply in the case of young people, whose neurological and cerebral development is not yet complete.
Though it is normal, as he stresses here, to be vigilant in such matters and to show concern about the consequences of over-exposure to screens, we must also keep matters in proportion: whether we are speaking of television, video games or social networks, it all depends on the contents and activities concerned. Some television programmes or video games may have harmful effects on health or brain development, but others may also stimulate and enhance the capacities of the children and adolescents who engage with them. It is difficult, then, to reach a cut-and-dried decision on the need or otherwise to avoid exposing children to screens, and we will probably have to wait some years to have sufficient perspective to determine the actual consequences of that exposure.
This third strand in the “Brain” series initiated by Futuribles in January 2019 is given over to human-machine interactions (screens, computers etc.) and, more particularly, to the impact of screens on brain development and function, especially among young people. Daniel Marcelli, a child psychiatrist, looks here at the excessive early exposure of very young children to screens and its consequences for their development.
He begins by recalling a number of clinical findings made by early childhood specialists — children living increasingly in their own bubbles and being increasingly withdrawn — which suggest that exposure of very young children to screens is likely to have serious consequences on the development of their language skills, emotions and capacity to interact with others (whether children or adults). Stressing the importance of interactions with adults and the real world, Marcelli describes the mechanisms of the child’s construction of relational and emotional capacities, mechanisms that cannot be acquired through the use of screens and the associated digital contents but require human relations. Hence his call for a genuine mobilization on the part of the public authorities to intensify research in this field, sensitize the population to the risks of such early, intensive exposure of very young children to screens, and develop teaching tools on the rational use of these interfaces within the family.
In January 2019 (Futuribles 428),we began a series on the human brain, with the first dossier devoted mainly to what the cognitive sciences and neurosciences teach us about our capacities for lifelong learning. A second dossier appeared in July 2019 (issue 431) on the plasticity of the brain, its anatomy and its capacity to modify its own functioning as it receives new experiences. This autumn, we open up a third strand on human-machine interactions and, more particularly, the impact of screens (TV, computers, mobile phones) on brain development and function, particularly among young people.
Jean-Pierre Bellier, who took part in producing this third strand, introduces the subject by reminding us how far human-machine interactions have now developed and evolved, and by stressing a number of questions inherent in these interactions which will be treated in greater detail in the other articles in this new dossier.
Après TV lobotomie , ouvrage dans lequel il prenait vigoureusement à partie des défenseurs du petit écran et leur déni quant à son emprise et ses effets dévastateurs sur l’intelligence humaine, après L’Anti-régime  puis L’Anti-régime au quotidien  dans lesquels il dénonçait l’aberration physiologique des régimes restrictifs, Michel Desmurget est de retour et nous invite à le rejoindre dans le cœur de cible des neurosciences cognitives, son ADN scientifique. Particulièrement argumenté, illustré par d’abondantes ...
(encore 1123 mots)
Following on from the first dossier on the human brain, devoted essentially to what the cognitive sciences and neurosciences teach us about our lifelong learning capacities (Futuribles 428, January-February 2019), we are publishing a second dossier on the plasticity of the brain, its ability to modify its operation over the course of its life as a function of experience.
Since this second dossier is more concerned with the morphology and operation of the brain, it relates more to the life sciences and is, for obvious reasons, introduced by Jean-Pierre Henry who has coordinated all the contributions. His article first explains how the brain is formed (in utero) from the sixth week after conception, then how it is shaped throughout existence and how it is permanently regenerating in the hippocampus. Henry then explains how the brain’s capacities develop — particularly those required for memory.
In the third part of his article, Jean-Pierre Henry sets out the pathologies the brain may suffer (particularly cerebrovascular accidents) and the progress made in their treatment. Going beyond this, with reference to the challenges that may confront the brain as it ages (particularly Alzheimer’s, Parkinson’s and ALS, which affect a large number of individuals), he shows how plasticity enables the brain to find solutions in cases of dysfunction, and the advances that may ensue from this in the understanding of these pathologies and the potential therapeutic responses to them.
Voici un numéro de Futuribles bien conforme à l’objectif que nous poursuivons : celui de pouvoir fournir à nos lecteurs une analyse rigoureuse et néanmoins lisible des grands enjeux du monde contemporain appréhendés au travers d’une démarche résolument pluridisciplinaire et prospective. Et comme Futuribles est une revue, non un magazine, point n’est question ici de « surfer » rapidement sur tous les sujets, aussi nous faut-il faire preuve de discernement dans la composition de chaque numéro. Alors que le précédent ...
(813 more words)
Continuing Jean-Pierre Henry’s argument, Hugues Duffau shows that the brain’s plasticity, its permanent capacity to reorganize itself to adapt to circumstances, ensues from the fact that it is an extremely complex system, all of whose parts function interactively, by processes into which he affords us some insights here. Accordingly, Duffau addresses the very widespread idea that each area of the brain corresponds to a given function (movement, language, memory, emotion etc.), a theory known as localizationism. That idea, he asserts, is refuted by the observation of the chain reactions that connect all the parts of the brain that become activated, bringing into play that synaptic plasticity so specific to the human brain, which, incidentally, distinguishes it fundamentally from what is called Artificial Intelligence.
The author, a famous neurosurgeon, is certainly well placed to show, for example, that the ablation of a cerebral lesion cannot be performed without respecting the entirety of the dynamic neuronal network that is unique to each person and constantly evolving — which, to reiterate, is not the case with machines. He reports on the progress achieved in the understanding of the anatomy and highly complex functions of the brain, since it is now possible to map these with ever-increasing accuracy and, where necessary, repair them. Lastly, he alerts us to the risks inherent in artificial neural networks — a pale copy, in his view, of human neuronal networks — which could ultimately result in a deterioration of the neuro-plasticity of the human brain.
Karim Benchenane’s article, the third contribution to our dossier on brain plasticity, revisits this notion to explain that it covers “the way experience will, in the long term, modify the effectiveness with which neurons are able to communicate among themselves.” Though he stresses the primordial role of the earliest years, he nonetheless observes that, contrary to what we had long thought, this plasticity lasts a whole lifetime.
Benchenane then explains the mechanisms governing memory, while stressing that there are different forms — most notably, short-term and long-term memory — not all of them dependent on the hippocampus. Then, as learning and memorization consume a large part of the energy required for brain functioning, he reminds us of the importance of sleep, which is required for physical recovery. He is careful also to underscore that these processes are still subjects of scientific controversy, which he lays out here very clearly and interestingly.
Lastly, Benchenane alerts us to the risk of falling victim, as we sleep, to social conditioning and he closes his article with some thoughts on brain-machine interactions, a subject to which we shall return in our third dossier on the brain, to be published at the end of the year.
Préfacé par Jean-Pierre Changeux qui alerte d’emblée le lecteur sur l’importance de comprendre ce qui distingue le cerveau humain de « la ferraille » computationnelle, l’essai d’Olivier Houdé est à n’en pas douter un pavé dans la mare : il renvoie à la fois dos à dos celles et ceux qui réduisent le développement de l’intelligence humaine à un processus adaptatif linéaire et / ou incrémental, et celles et ceux qui ne considèrent le cerveau que comme une ...
(encore 978 mots)
Attracteur étrange, le cerveau est depuis ces dernières années objet de toutes les attentions, au risque d’ailleurs de nombreuses redondances si ce ne se sont approximations voire fantasmagories. Catherine Vidal, neurobiologiste et directrice de recherche honoraire à l’Institut Pasteur de Paris, tente d’éviter cet écueil et nous livre dans un bref essai son analyse sur les apports comme sur les dangers des dernières innovations technologiques issues de la recherche en neurosciences. Son propos vise à modérer l ...
(564 more words)
Dans ce livre, Luc Julia nous parle de lui-même, de l’intelligence artificielle à laquelle il a consacré beaucoup de sa vie et de ce qu’elle pourrait faire à l’avenir pour être utile aux vraies gens. Et, loin des fantasmes sur une intelligence artificielle qui nous dominerait, il nous livre son expérience, la chronologie des avancées du numérique, la personnalité de ses acteurs et les faits en matière de technologie qui constituent autant de clefs pour comprendre ce ...
(974 more words)
J’adresse mes vœux les meilleurs à nos lectrices et à nos lecteurs pour qu’ils puissent, durant l’année qui vient, transformer eux-mêmes le souhaitable en probable, donc être acteurs plus qu’esclaves de leur avenir tant individuel que collectif. L’an dernier déjà , je les invitais à devenir « le changement qu’ils veulent voir dans le monde », à ne point attendre donc d’instances supérieures qu’elles accomplissent seules les réformes qui s’imposent, notamment parce que ...
(872 more words)
Also as part of our special dossier on research into the brain and learning processes, a major question is raised in this article: are recent advances in artificial intelligence (AI), and particularly the rise of neural networks, liable to put in doubt the supremacy of the human brain? What differences in nature, what conflicts or complementarities are there between these two forms of intelligence?
After reminding us of the birth of the neural networks field, the advances made with such networks and their recent successes, Jean-Claude Heudin lays out their limitations. He goes on to explain the specificity of neural networks and AI, which, he writes, “are not complex systems, but ordered systems” that may have superior capacities to humans with respect to certain tasks. By contrast, human intelligence is “many-faceted, emotional and empathic”; for that reason, it has superior abilities to AI when it comes to performing many other tasks and functioning in a complex environment. Lastly, taking pains to demonstrate the different forms of intelligence, Heudin concludes that AI and human intelligence are complementary.
The human brain, a vital organ unrivalled throughout the animal kingdom in terms of its performance, just keeps on surprising us. We have made considerable progress in recent decades in understanding its functioning. There is, however, still much to discover and explore, not only to understand how humans grow, reflect, think, learn, adapt, and feel, but to make advances in improving the brain’s functioning (particularly in terms of learning) and repairing its dysfunctions, whether mechanical, genetic, pathological etc. This is why Futuribles has decided to begin 2019 with an issue devoted very broadly to the human brain, the first in a series that will be completed over the coming months.
This article by Grégoire Borst makes a start by presenting the reader with an account of how the human brain operates. What do we know today of the way the brain is constructed from the embryonic stage through to adulthood? What is the pace of that construction? How is the neurological system established? What is the general architecture of the brain and to what specific aptitudes do its various elements relate? Does the brain have particular capacities for adaptation? Is it influenced, in its construction and operation, by the familial or socio-economic environment etc.? These are the big questions explored here, in order to understand the operation of that organ in an age when competition between man and machine is a topic of increasing concern, particularly in relation to advances in artificial intelligence.
Completing the series on the brain begun in this issue of Futuribles, this article by Jean-Luc Berthier shows concretely how the neurosciences can help to bring about advances in educational methods. Drawing on a number of experiments undertaken in France, particularly within the context of so-called cogni’classes, Berthier demonstrates how research in the neurosciences, by enabling us better to understand brain functionalities, offers teachers and their pupils new paths for learning.
For example, Berthier presents a whole series of new educational methods aimed at facilitating memorization, gaining pupils’ attention, and differentiating practices to meet the profile of the learner etc. He also describes the main steps that must be undertaken to construct a teaching project based on the cognitive sciences, as well as the pedagogical pathways that are most employed to this end. Lastly, he details the possibilities offered by the use of artificial intelligence in educational practices, while reminding us that the aim is to facilitate learning and the work of teachers, not to substitute for them. The new educational pathways opened up by the entry of the neurosciences into the classroom are encouraging, but require the training of specialist teams and the involvement of all the actors in the system (teachers, pupils, management) – which is asking a great deal in France!
A vital organ with levels of performance unrivalled among other living species, the human brain is a constant source of surprises. Substantial advances have been made in understanding its operation in recent decades. There is, however, much still to discover and explore, if we are to understand how human beings grow, reflect, think, learn, adapt, feel etc., and to advance in the ways of improving brain function (particularly in terms of learning) and repairing dysfunctions of whatever kind. This is why Futuribles has decided to begin 2019 with an issue devoted very largely to the human brain and, more specially, to advances in the cognitive sciences and neurosciences and their contributions to education and learning.
This article by Olivier Houdé lays out the – recent and doubtless still too limited – contributions of research in neurosciences and cognitive sciences to the field of education. Houdé particularly the stresses the importance of brain research, thanks to the observations of its in vivo functioning, for the understanding of children’s learning mechanisms. He puts especial emphasis on the two complementary forms of neurocognitive learning that are automation and control by inhibition (or “de-automation”). He explains how these show up in the brain and the thought systems that activate them. The advances in the understanding of these mechanisms have thus opened up new pathways in educational sciences.
Still as part of the series on the brain begun in this issue of Futuribles, Pascale Toscani, in the introduction to her article, raises a familiar question: why do we have to think before replying to a question that is put to us? The answer, she explains, is that “our brain works before us, before the information arrives in our consciousness,” because it is endowed with a capacity for anticipation based on everything it has registered in the past. But trawling our memories is not enough and the author shows us the frequent mismatch between question and response depending on the terms employed, on each person’s representations, on our cultural referents etc.
With the help of copious examples, Toscani alerts us to our cognitive biases, a deceptive, falsely logical from of thinking, which she aims, in part, to explain – first by describing in detail how babies’ brains are organized and operate and, second, how academic learning works. And, last, by taking up and delving deeper into the two systems of thought highlighted by Daniel Kahneman (automatic thought based on “procedural memory” and, by contrast, demanding reflective mental activity), in order to show the essential role intelligence plays in freeing us from received ideas and cognitive biases and enabling us to manage situations characterized by “cognitive dissonance”.
A vital organ with levels of performance so far unrivalled among other living species, the human brain is a constant source of surprises. Substantial advances have been made in understanding its operation in recent decades. There is, however, much still to discover and explore, if we are to understand how human beings grow, reflect, think, learn, adapt, feel etc., and to advance in the ways of improving brain function and repairing dysfunctions of whatever kind. This is why Futuribles has decided to begin 2019 with an issue initiating a series devoted to the human brain.
Jean-Pierre Bellier, who contributed in large measure to the production of this series of articles, lays out the reasons here that led us to take an interest in the subject: the technical advances in recent decades that have enabled us to observe and understand the organ better, and those who have opened up new perspectives in terms of brain research, if not indeed of a hybridization of human and artificial intelligence. He particularly stresses the issues inherent in the progress of the cognitive sciences and neurosciences, and their contributions in terms of education and learning which form the core of this first raft of contributions.
In the special dossier we are devoting in this issue to advances in research on the human brain and the contribution of that research to the development of educational practices – and, more generally, our capacities for learning – this article by Elena Pasquinelli is important for providing an opportune reminder that, though that research may contribute greatly to improving educational practices, we should remain vigilant towards so-called discoveries that are, in reality, unfounded.
After stressing the legitimate hopes that may be vested in this research, Pasquinelli warns us against false – or perhaps merely fashionable – beliefs in such things as “the Mozart Effect” (the idea that listening to classical music might improve our intelligence). She gives many examples of these “neuromyths”, as she calls them: widely believed theories that are not supported by any serious research. Ultimately, her article invites us to maintain a discerning stance on these matters, particularly where the links between science and teaching are concerned.
Entre « pari de Pascal » et interprétations à la « mode Trump », Vincent Berthet nous entraîne dans l’univers de l’irrationalité des comportements humains et de ce qui en constitue une des caractéristiques, à savoir ces biais cognitifs dont nous sommes toutes et tous victimes consentantes parce qu’inconscientes. L’auteur s’essaie dans un premier temps — pour mieux la réfuter ensuite — à l’exploration de la théorie du « choix rationnel » en tant que principe unificateur capable de rendre compte du ...
(1186 more words)
Dans cet ouvrage, Leslie Valiant, professeur d’informatique théorique et de mathématiques appliquées à Harvard, nous offre une approche originale et ambitieuse des algorithmes, pièce maîtresse de l’intelligence artificielle. Le texte s’ouvre sur une référence à Von Neumann qui affirmait, en 1947, que « si les gens ne croient pas que les mathématiques sont simples, c’est parce qu’ils ne mesurent pas à quel point la réalité est compliquée ». À la fois un paradoxe et une évidence. La ...
(1011 more words)
Ouvrage collectif associant une trentaine de contributeurs des plus réputés en sciences cognitives — auteurs ou coauteurs pour l’occasion de 12 chapitres et de 12 focus —, Le Cerveau et les apprentissages s’adresse à celles et ceux qui cherchent à comprendre comment et pourquoi la sortie, relativement récente, de la confidentialité des recherches in vivo sur le cerveau bouscule à ce point les poncifs et stéréotypes jusqu’à parfois proposer « la bonne façon » d’apprendre, donc d’enseigner. À l ...
(929 more words)