False dilemmas and real challenges: GenAI and the algorithmization of experience

Abstract

Generative artificial intelligence (GenAI), which emerged just a few years ago with ChatGPT at the forefront, is causing an epistemological shock. This, in turn, is translating into transformations across the social, cultural, political, and legal spheres. Most analyses cast GenAI in one of two roles: either as a nearly magical tool that will solve all problems with speed and efficiency, or as the destructive tool that will erase creativity, imagination, and ultimately the humanity in what we create. In any case, what we should not do is fall into the trap of competing with large language models (LLMs). To do so would be to assume that human intelligence is an accumulation of information, and that creativity can be reduced to the most probable response. What we are looking for is not always found on the internet; often, it must be invented. “A plausible way out of this debate is to acknowledge the paradox that current artificial intelligence is simultaneously very intelligent and very stupid; its stupidity lies in the fact that, when it makes an intelligent decision, it has no way of knowing it (Dessalles 2019). What we would have, then, are “idiot digital savants” (Carr 2014; Domingos 2015)” (Innerarity, 2025, p. 37, my translation).