Abstract

The development of artificial intelligence has proceeded within a disciplinary culture whose positivist epistemological foundations generate a constitutive blind spot: the systematic exclusion of tacit, embodied, relational and contextually situated knowledge from what counts as knowledge at all. This article argues that this exclusion is not a technical limitation, but a structural condition of AI systems as currently built, with serious consequences for individuals, institutions and the social order. The method is interdisciplinary and critically synthetic, integrating the anthropology of science and technology, the philosophy of language and the literary-philosophical tradition of techno-critique. Three original contributions are advanced. First, the Faustian framework is reformulated as a collective rather than individual pact: the AI contract implicates a civilisation. Second, the prevalent reductionist critique is refined: the error is not quantification of incommensurable goods but the enforcement of total orderings upon value landscapes that admit only partial orderings. Third, the stochastic parrot objection is reconciled with the Faustian analysis through the concept of institutional amplification: AI’s danger resides not in its intelligence but in the authority conferred upon its incomprehension. The article concludes that Mephistopheles, not Faustus, is the more precise figure for AI itself, and asks whether the defunding of humanities disciplines, those best placed to navigate the moral challenges AI presents, constitutes the most consequential characteristic deletion of all.

IPC Classification

Keywords