Can AI Produce Meaning?
Epistemology & Philosophy of Science · 2026 · 63 (2) · 154-171
DOI: 10.5840/eps202663230 ↗Kurz gefasst
Fachartikel in der Zeitschrift Epistemology & Philosophy of Science, 2026. Verfasst von Asya A. Filatova.
Abstract (Originalsprache)
The rapid development and widespread deployment of artificial intelligence (hereafter, AI), particularly in the form of large language models, has prompted a renewed wave of philosophical debates concerning the semantic competences of technical systems and the criteria for evaluating the outputs they generate, especially textual outputs. This article examines a set of central problem areas and their treatment in contemporary philosophical literature. Drawing on the theory of L. Floridi, the paper first raises the question of whether AI systems can possess and generate semantic capital (hereafter, SC), understood as any content that enhances an agent’s capacity for semantization. The paper argus that Floridi’s account remains largely anthropocentric, assigning neural networks the role of a “semantic broker”, that is, an entity capable of increasing SC through computational capacities while lacking meaning and content in the strict sense. The paper then analyzes the “stochastic parrot” problem within the framework of competing metasemantic theories. It shows that the same conceptual resources function both to deny and to support the claim that AI systems can produce meaning, thereby revealing a general limitation of anthropocentric approaches. Using the method of “anthropocentric abstraction”, proposed by H. Cappelen and J. Dever, the paper illustrates how Kripkean externalist semantics can be applied to the analysis of AI-based decision-making systems. Particular attention attracts the problem of alien content generated by neural networks and to the challenges it poses for human understanding; the paper outlines several scenarios for bridging the communicative gap between humans and AI systems. The conclusion argues for a symmetric approach to evaluating the semantic competences of humans and AI systems, taking into account differences in their internal organization and modes of operation. It further emphasizes the cognitive co-evolution of humans and AI systems, as well as the importance of investigating the social mechanisms that sustain SC.
Zitieren
APA
Filatova, A. A. (2026). Can AI Produce Meaning?. Epistemology & Philosophy of Science, 63(2), 154-171. https://doi.org/10.5840/eps202663230
Vancouver
Filatova AA. Can AI Produce Meaning?. Epistemology & Philosophy of Science. 2026;63(2):154-171. doi:10.5840/eps202663230
BibTeX
@article{Filatova2026,
author = {Filatova, Asya A.},
title = {Can AI Produce Meaning?},
journal = {Epistemology & Philosophy of Science},
year = {2026},
volume = {63},
number = {2},
pages = {154--171},
doi = {10.5840/eps202663230}
}
EUVIMED