The recent release of the “chatbot”, ChatGPT, developed by the OpenAI company, immediately triggered an avalanche of comments. What was most highlighted is that the system, designed to answer written questions, offers responses of “surprising” syntactic quality and consistency.
However, as if asking for more, demonstrating a lack of critical thinking, this observation was tempered by the fact that the result is still in its infancy; a certain number of advances are necessary to equal a human being. But this is precisely where our great illusion lies. The one that consists of believing that these are devices that use a language similar to ours.
That is why it is necessary to go and see – far from the discourse forged by the digital industry, which we so often take literally – what are the springs of these technologies. These are capable of dissecting the structure of textual corpora available in databases or on the Internet to elaborate semantic laws.
In this sense, what characterizes these statements is that they are nothing more than the production of algorithms that feed on statistical analysis, thus taking already existing records as their only source. In this case, they are alien to what the so-called “natural” language supposes.
Because the specificity of human language lies in the fact that it arises from a tension between a vast lexicon, made up of words, grammatical rules, and our ability to generate formulas in a relationship with time that is not that of an exclusive attachment to the past, but that of a dynamic conjugated with the present and in constant evolution.
When we speak or write, we constantly drink from an ocean of phraseology, indeterminately adjusting to a specific context each time. Any locution, written or spoken, comes from an emergence that invariably exceeds any previous schematization. And it is this dimension that is absent from the mechanical verb, because it is only the result of parameterizations that only respond to certain functionalities. It is the one that operates, for example, in the personal assistant Siri (owned by Apple) that tells us “what can I do for you? “. Or in connected speakers like Alexa, developed by Amazon, whose sole purpose is to guide our decisions for primarily business purposes.
More than naively wondering if these systems will soon replace us in the writing of texts – then a sign of a definitive renunciation of the use of one’s own reason – can we perhaps see the civilizing model that is quietly being instituted? The one that results from a double transformation of our relationship with language. One will see, on the one hand, the so-called “generative” artificial intelligences endowed with the power of speech with supposedly identical airs to ours, to which little by little the care of managing our relationships with others will be delegated, as well as many others of our current tasks. This, according to a phenomenon that exempts us from this faculty that, however, conditions our right to pronounce ourselves in the first person and to conduct ourselves according to our judgment, within a free and plural society.
On the other hand – and this is the main objective – technologies are erected that lavish us with their good words with a familiar and intimate tone, but considered superior and awake, inciting us to act in this or that way, within an instrumentalized dialogic relationship. It is a model that has been in force for fifteen years, due to the uninterrupted development of artificial intelligence.
It is the model for the robotized interpretation and recommendation of our gestures for our supposed greater comfort. As such, not a “surveillance” capitalism, but more exactly “the management of our welfare” now prevails. In which we continue to hunker down, holding these digital spectra for highly enlightened entities and guiding us day and night on the right path.
What should also be noted is that these devices are constantly becoming more sophisticated, in particular due to automatic learning (self-learning processes), called to adopt more and more natural appearances and from which it will be difficult to deviate from their exhortations that seem to come from omniscient consciences. . It is an industrialized language taking our place, which guiding us will continually become a habitus, especially with the younger generations, who quickly find these uses so easy and self-evident.
It is also an epigenetic process that is underway, so that our mind, which is built to be fully active, will let itself be carried away by these systems, gradually reducing the use of our expressive faculties. Because it is from a medium and long-term perspective that we must assess the scope of the consequences induced by the development of these technologies.
These gigantic convulsions -ultimately anthropological in scope- that we have been experiencing without truce for twenty years, are above all the result of the digital industry. This is the case of Elon Musk, creator of Open IA, now owned by Microsoft. But they are also companies that, all over the world, are strongly engaged in these fields of research.
“Who is speaking?” Nietzsche asked. Are these programs whose language, deprived of its vital dynamics, conveys a vision of the world based on generalized reductionism and utilitarianism, do it today? Or, on the contrary, are our voices, each one unique, arising from our spirit and our sensitivity, the only ones capable of establishing active links with others and with reality?
More than ever, the time has come to raise the question of language, the one we want to speak -in our name and in a truly common whole- as the moral, political and civilizing question -literally the first- of our time.
* Éric Sadin is a philosopher, specialist in the digital world. Last published book: The era of the individual tyrant: the end of a common world (Caja Negra Editora, 2022).
Traducción:Telita Doyhambéhère