AI Art: The Semiotic Error as the Manifestation of Artificial Intelligence
AI · 03.11.2026 · 7 (11) · 454
DOI: 10.3390/ai7110454 ↗Kurz gefasst
Fachartikel in der Zeitschrift AI, online erschienen am 03.11.2026. Verfasst von Daria Bylieva. Die Arbeit nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This paper contends that the semiotic failure in generative art is not merely a technical anomaly to be optimized away, but a critical aesthetic event that exposes the internal mathematical constraints of non-human computational models. Bridging Donald Campbell’s blind-variation model with Viktor Shklovsky’s ostranenie and the Heideggerian phenomenology of tool-being, the study analyzes algorithmic error as a disruption that shifts generative systems from a state of transparent functionality into a mode of confrontational presence. Utilizing a methodology of technical–aesthetic reverse-engineering, a spectrum of contemporary artistic practices is examined to demonstrate how structural boundaries become legible when human linguistic codes fracture against the stochastic reality of deep neural networks. The research establishes a four-tier operational typology that conceptualizes the boundaries between normal stochastic variation, machine hallucination, technical glitches, and artistically induced errors. The analytical framework demonstrates that the authentic disclosure of the non-human parameters of artificial intelligence occurs not within the regimes of normative prompt execution, but through the tactical subversion and deactivation of the built-in optimization boundaries and platform filters engineered to constrain the algorithm. Within this critical threshold, visual anomalies offer a plausible structural interpretation of latent clipping failures, feature entanglement, and motion-vector degradation occurring when hyper-amplified Classifier-Free Guidance scales collide with non-congruent data matrices. Ultimately, the embrace of computational failure reconfigures the AI error from a technical deficit into an interpretive threshold, serving as a phenomenological index of operational friction that exposes the fragile, non-semantic structure of deep neural networks.
Zitieren
APA
Bylieva, D. (2026). AI Art: The Semiotic Error as the Manifestation of Artificial Intelligence. AI, 7(11), 454. https://doi.org/10.3390/ai7110454
Vancouver
Bylieva D. AI Art: The Semiotic Error as the Manifestation of Artificial Intelligence. AI. 2026;7(11):454. doi:10.3390/ai7110454
BibTeX
@article{Bylieva2026,
author = {Bylieva, Daria},
title = {AI Art: The Semiotic Error as the Manifestation of Artificial Intelligence},
journal = {AI},
year = {2026},
volume = {7},
number = {11},
pages = {454},
doi = {10.3390/ai7110454}
}
EUVIMED