Chatgpt: inteligência artificial, pensamento e linguagem em questão

Abstract

It is often the case, even among scholars, that ChatGPT (as well as other Ais) is attributed anthropomorphic characteristics. The use of expressions such as reasoning, understanding, intelligence, hallucination, thought and even consciousness is not, in some cases, reduced to a metaphorical function. In this essay, we tackle the assumption that the machine, when executing AI algorithms, is capable of thinking or speaking, just like a human being. The goal is to address this “badly analyzed composite” (DELEUZE, 2012), in which different natures or essences are confused. In short, with ChatGPT as a backdrop, the aim is to revisit the old questions: “can machines think?” and “can machines speak?”. To this end, ChatGPT is first presented technically, without forgetting to situate it through some political and social crossings. Two major themes are then addressed – Thought and Language – with the aim of outlining the differences in nature between technical machines and living machines, between ChatGPT and human beings. In doing só, we seek to highlight the implications of different images of thought and language. Finally, ChatGPT’s limitations are highlighted, such as its inability to deal with the events of the present.