The Cartesian Paradox: Thinking AI and Non-Thinking Humans?

Abstract

The rise of artificial intelligence has revived a fundamentally Cartesian question: what conditions must hold for something to count as genuinely thinking? This paper examines the relationship between cognitive delegation and the Cartesian cogito, arguing that contemporary AI systems challenge not the logical validity of the cogito itself, but its cultural and epistemic centrality. Drawing on a methodological distinction between functional, Cartesian, and agential forms of thinking, the paper develops what is termed the "Cartesian paradox": while humans increasingly delegate judgment and reasoning to AI without necessarily ceasing to exist, machines increasingly reproduce the external markers of intelligence without thereby becoming res cogitantes. The analysis shows that the cogito guarantees certainty of existence through first-person self-presence, but does not imply that the suspension of reflective thought entails ontological annihilation. At the same time, the paper argues that AI systems, despite their growing functional sophistication, remain forms of res extensa insofar as they lack subjective self-awareness and agential self-relation. The conclusion is that the real danger posed by AI is not the replacement of human beings by conscious machines, but the progressive reduction of thinking to operational performance, externalized computation, and algorithmic delegation.