Artificial Intelligence and Understanding. On the Artificial Way of Considering Human Understanding

Abstract

Through various philosophical contributions–the Heideggerian concepts of the world (Welt) and the they (das Man), Merleau-Ponty’s perception, the background of Hubert Dreyfus and Charles Taylor, and Wittgenstein’s Lebensform –this work aims to contrast a phenomenological-ontological framework of understanding-interpretation with the traditional epistemological-anthropological model, which considers the human being as a knowing subject. Based on this, two paradigms of Artificial Intelligence (AI)–the symbolic model and the neural statistical model–are compared with these two ways of understanding engagement with reality, following the reflections of Hubert Dreyfus. Finally, the paradigm of working with the language patterns of LLM (Large Language Model) programs is presented as a diverse alternative within the neural statistical paradigm, drawing on a conception of language from the later Heidegger, after Being and Time.