Twice a month, The duty challenges enthusiasts of philosophy and the history of ideas to decipher a topical issue based on the theses of a prominent thinker.
While “ChatGPT [possède] the record for the fastest growth in the number of followers”, the dispersion of AIs in all social and economic spheres cannot now be stopped. Some even argue that we are a few years away from a first singularity: a technological development so rapid and radical that it becomes impossible to predict its impacts. Why do these radical innovations continue their rampant deployment, to the point where thousands of researchers and entrepreneurs have called for a halt to AI research and development in order to look into “protocols and rules to make future advances more responsible and transparent”? Do engineers, programmers, designers, mathematicians live in another world not to worry and pursue this race? For Boltanski and Thévenot and their economies of grandeur, the answer is: yes, they live in another world.
Justify yourself to remain “great”
This world is the “industrial world”, one of the common worlds proposed by Boltanski and Thévenot. Six in number – merchant, industrial, civic, domestic, inspired and opinion – these worlds exist and are maintained thanks to the actions, the words, the individuals and the objects organized around what the authors call the “principles common superiors”: rules constantly present, reiterated and to which the actors of each world refer. In conflict or in the face of criticism, individuals justify themselves by these principles, making them even more solid and important. To live up to these great principles specific to their world, the actors invest actions, speeches and objects to achieve what the authors call the test of greatness. Hence the title of the book: Of justification. The economies of scale (Gallimard, 1991).
The industrial world, encompassing science, technology, engineering and medicine, “is thus [fondé] in the objectivity of things that form naturally”. There are considered “great” individuals who “work to discover and coordinate the general facts suitable to serve as a basis for all combinations of culture, commerce and manufacture”. In the industrial world, the test of size is achieved by demonstrating the effectiveness and performance of the technical object or the scientific method. Thus, members of the industrial world (AIs) maintain themselves in a state of greatness by arranging massive data, algorithmic models and machines capable of processing and digesting astronomical masses of data and queries to answer all questions, in all fields, with an efficiency unequaled by man.
For Boltanski and Thévenot, the six common worlds are not ideologies, ways of thinking or dogmas. They are arrangements of actions, of people, of objects, of words which become so recurrent, so solid that “the figure of the common superior is reality”. In other words, the daily life of the actors of a world finds itself so entangled in the common superior principle, the multiple tests of magnitude and the arrangements of this world that they cannot escape the test of magnitude, at the risk of to be in “failure” and to become “small”: “Weighs on them the worry of seeing the principle from which they derive the share of greatness from which they can benefit, however small, collapse and of throwing to down the very order of things. »
Clash between sizes
This tangle of investments and hardships that ensure the greatness of a world does not make its actors insensitive or blind to other worlds. It is moreover in the clash between the magnitudes of two worlds that conflicts and criticisms emerge. Let’s take the fears raised from the “inspired world”, that of art, creativity, inspiration that brings out “the unusual, the spontaneous, the bizarre” and appropriate emotions. In this world, there is recognized as great anything that “subtracts from measure”, which “springs from inspiration”, which is experienced in “an inner experience” and which transforms individuals. We easily see the fundamental conflict between the two worlds: the efficiency and performance of a methodical routine versus the spontaneous, unpredictable outburst that is experienced without measure.
The discomfort raised by the interference of generative AIs is understandable because they make their way by monopolizing the test of magnitude of the inspired world, without however going through inspired forms of investment. The stunning productions of AIs in art, music, literature have generated “emotions and passions” specific to the inspired world, “experienced as devouring, frightening, enriching, exciting, exalting, fascinating, disturbing, etc. without, however, being “worthy” of the investment specific to creation: “love, passion, anxiety, doubt, the desire to create”. Some could argue that the artistic greatness of AIs is not one, that the works produced are statistically conventional and that they will never manage to produce this “unusual”, “unspeakable” part of the works. human. This is to forget that this incapacity constitutes, in the words of the authors, a failure: “The failure of people manifests itself when they are not up to the task, when they do not highlight objects to the best of their sizes. and that they therefore did not make the sacrifice supposed by their apparent state of greatness. The creative mimicry of AI is the new test of magnitude of the industrial world, which actors must pass or risk becoming “small”.
Device for the common good
During the clash between two worlds, Boltanski and Thévenot remind us that actors can a) confirm what is great with the test specific to their world (ex. show that an AI can create a painting so effectively that it is sold at auction); b) to arrive at a temporary arrangement between the worlds (the introduction of an invisible signature in the texts produced by ChatGPT); or c) create compromise devices, which we will focus on here given their potential sustainability. This device must “place at the service of the common good objects composed of elements from different worlds and endow them with their own identity, so that their form is no longer recognizable if one or other of the elements of disparate origin of which they are made up. The Montreal Declaration for Responsible Development of Artificial Intelligence, which calls for the development of an “ethical […] inclusive, equitable and ecologically sustainable of AI” is an example. However, for a device like this to be able to operate in the labs and the start-ups, there must be actors there – artists and people from the human and social sciences – who can ensure the maintenance of the new greatness common to the two worlds. and work for the common good. Without a collective oversight of the new common test (an efficient and socially responsible AI), the test of the dominant world – the efficiency of the industrial world – will return at a gallop.
The social experiment of creativity
But what should this new hybrid test of magnitude be to the industrial and inspired worlds? According to the authors, the compromise device will have to ensure the maintenance of our relationship to the human and social experience of creation: human, materials combined by AI; social, in our relationship to the processes leading to these materials.
Faced with a work, the device will have to clearly identify a #MadeByAnAI work so that we can test it (literally, “to do an experiment likely to establish the value of something”) for what it is: a statistical combination that is not made of these “detours, sufferings, experiences, doubts, chances” of a human work and of our attachment or our resonance to these experiences. In the same way, conversational robots should allow us to go back to the source, as many artists claim, to understand the still human relationships (at least for the moment) of the materials mobilized during the “creation”. On the one hand, we will be (hopefully) outraged when we reach the threshold where the productions we consume will be AIs that are inspired by other AIs. On the other hand, we will be able to separate the human act from the machine act. Better, we can develop a digital literacy that allows us to discover authors, artists and materials at the origin of artificial production.
In short, it is essential not to confuse the creative production of humans with that of machines, or let the part of humanity disappear into the black box of artificial productions, otherwise the “distinctive dignity of humanity is threatened by the treatment of people like [s’ils étaient] things. The greatness of the objects and devices created can be confused with its dignity to the point of blurring the limit of humanity”. The production of an AI is a technical object. Human creation is a social process. The meeting of the first must take place in the recognition of the second.
Suggestions ? Write to Robert Dutrisac: [email protected].