In a famous 2008 ‘End of Theory’ article, Chris Anderson provocatively suggested that in the era of massive data collection, the scientific method, and, in particular, the theorizing activity that ‘it implies, would have become obsolete in the face of the computing power of algorithms. And yet, if it is obviously mathematics, and in particular logic, computer science and algorithmics, which have enabled the development of artificial intelligence systems, it is clear that despite their increasingly impressive performance , there is indeed an activity that the said “artificial intelligences” do not manage to simulate: the invention of a mathematical theory remains for them something completely foreign.
For the 20 years of Futura,, Entrepreneur and Co-President of the National Digital Council, joins the editorial staff to offer you throughout this special day, topics that are agitating the digital community. The articles that you will discover today aim to take a step aside, to raise different questions and to propose a meeting between science and society on ineluctable subjects around digital technology.
As the philosopher and physician wrote in 1980, in an article entitled : “One thing is the calculation or processing of data according to instructions, something other than the invention of a . Calculate the trajectory of a spatial is under the control of the computer. Formulating the law of universal attraction is not a performance. No invention without awareness of a logical void, without tension towards a possible, without the risk of being wrong. “
From program to invention: calculating is not thinking
The invention, Canguilhem recalls, corresponds to a ” thought situation where we aim for what we do not see », And which, by definition, cannot therefore be programmed. ” To invent, he still maintains, is to create information, to disrupt habits of thought, the stationary state of knowledge »: This does not therefore amount to executing a set of predetermined rules or to relating data, but to engender novelty in relation to what is already there, in a way that is unpredictable for the inventor himself, who is often the first to be surprised by his discovery or his theory.
Many will argue that since the 1980s, “machines” have been transformed, that they no longer operate under the constraint of determined programs, that they are capable of “learning” based on the data collected, that we today we are dealing with “learning machines”, even “moral machines” or “intelligent machines”. But on closer inspection, Canguilhem’s argument perhaps deserves a little more attention. Because while technologies have of course evolved, the discourse that accompanies them has ultimately changed relatively little. Already in 1980, Canguilhem reported “ abuse of irrelevant expressions such as “conscious brain”, “conscious machine”, “artificial brain” or “artificial intelligence” », Which are all based on aCome in , the machine and the mind.
From artificial intelligence to collective intelligence: digital technology at the service of knowledge
However, for Canguilhem, whatever the performance of the technical devices, such analogies remain problematic: if certain cerebral processes or certain intellectual operations could have served as a model forartificial machines, this does not imply that these machines, once constructed, can serve as a model for understanding the or to be attributed intellectual faculties, even though it is precisely this thought or these faculties which allowed their conception and their construction, through the communication of the many brains which then joined forces to develop them.
Indeed, far from being reduced to a set of neural processes or formalized logical operations, thought constitutes what the philosopher described as a process of psychic and collective individuation, that is to say, a process during which living individuals collectively relate to each other by sharing symbols and meanings, which are always inscribed, deposited, sedimented in material supports (printed, , digital) allowing to manipulate them. Thought is therefore neither in brains nor in machines, but rather Come in brains, in symbolic environments shared by individuals and supported by all kinds of technical devices, which in turn affect cerebral functions, psychic capacities and collective knowledge.
Therefore, in the age of “artificial intelligence”, the question that arises is not whether machines can think, but how can automated computing devices and digital publication media transform symbolic environments and shared knowledge. It is on this question that the members of the National Digital Council asked themselves, in a note entitled ” “. From computerization to . Indeed, if the computerization of societies has continued to develop, if technologies have continued to improve, the question of the empowerment of citizens now seems to arise: if we have all become users of the digital, it is not certain that we understand the functioning of the technologies we use, nor the anthropological stakes of this revolution.
From computerization to capacitation: towards a shared digital culture?
Compared to other technical revolutions, the digital transformation has occurred at aunprecedented, and digital devices spread in societies long before the knowledge necessary for their taming could develop, so that users often find themselves in the position of consumers of services or “addicts” to services. . However, the indiscriminate use of a technology can be a factor of “incapacitation”, insofar as technical circles configure our uses, shape our behavior, and deeply affect our faculties as well as our relationships. Technologies are not mere means or instruments: they transform us as much as we use them. If we do not understand the devices that surround us and on which we depend on daily basis, we risk finding ourselves enslaved to automatisms that we have not chosen and of which we are not aware.
To face this risk, the need for a digital culture seems to be essential. This culture is not limited to the acquisition of technical skills, but also implies a historical awareness of the evolution of technologies, enabling citizens, and in particular the younger generations, not only to use them, but also to understand and understand them. to transform them. Provided that they are designed and developed in this sense, digital media contain unprecedented potential for the co-construction of knowledge, argued controversies and the sharing of knowledge: unlike audiovisual media which imply a structural division between producers and receivers of content, digital technology allows receivers to move into the position of producers and contributors, on condition that they have the practical capacities and critical reflexivity allowing them to appropriate the technical devices.
The development of a shared digital culture would thus make it possible to inventand to the singular desires of the populations: far from promoting the “end of the theory” in the face of algorithmic calculations, it would be a question of putting at the service of collective intelligence.
All the articles of the special day: