The rapid development of Artificial Intelligence has transformed the global technological landscape within only a few years. Large Language Models, automated translation systems, predictive algorithms, and generative AI are now becoming central to education, communication, administration, research, and economic production. Yet behind this technological acceleration lies a major question that remains insufficiently explored: can Artificial Intelligence genuinely function in a multilingual world without reinforcing new forms of linguistic domination? This question becomes particularly urgent for the Global South, where linguistic diversity is not an exception but the normal condition of social life. Thousands of languages continue to exist across Asia, Africa, and Latin America, often within multilingual societies where several languages, dialects, and religious traditions coexist simultaneously. Modern AI systems, however, are largely concentrated on a very small number of dominant languages, supported by massive amounts of digital data. This imbalance raises the possibility that AI may unintentionally create a new form of digital hierarchy with thousands of minority languages reduced to translated extensions of the few dominant ones. The problem is not simply technological. It is also cultural, linguistic, philosophical, and civilizational. After Jean Coulardeau (1942-2022) and his book, The Computer, The Ultimate Tower of Babel, Éditions La Galipote, 2006, Pope Leo XIV just required the slowing down of Artificial Intelligence to avoid a Tower-of-Babel syndrome, and to remain human, in his encyclical letter, MAGNIFICA HUMANITAS on safeguarding the human person in the time of artificial intelligence. Yes, indeed, our times are recurrently disquieting.
COULARDEAU et al. (Sat,) studied this question.