Artificial Intelligence and Security in the Age of Realfakes
Facta Simonidis · 30.06.2026 · 19 (1) · 133-168
DOI: 10.56583/fs.2927 ↗Kurz gefasst
Fachartikel in der Zeitschrift Facta Simonidis, online erschienen am 30.06.2026. Verfasst von Mikołaj Kabała. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).
Abstract (Originalsprache)
his article analyzes the significance of security in the context of AI-generated images and clarifies the concept of “realfake” as a category describing synthetic or hybrid visuals that mimic the evidentiary power of documentary imagery. In the study, 217 participants evaluated 18 stimuli – real photographs, DALL-E images, and images stylized in the Ghibli aesthetic – in terms of authenticity, sharing propensity, and perceived risk. A repeated-measures ANOVA with Greenhouse–Geisser correction revealed significant differences between conditions. Real photographs were rated as the most authentic, DALL-E images as the least authentic, and Ghiblistyle images generated the highest sense of uncertainty and risk. The results justify combining detection, provenance solutions, and media literacy education focused not only on deception but also on the destabilization of the evidentiary environment.
Zitieren
APA
Kabała, M. (2026). Artificial Intelligence and Security in the Age of Realfakes. Facta Simonidis, 19(1), 133-168. https://doi.org/10.56583/fs.2927
Vancouver
Kabała M. Artificial Intelligence and Security in the Age of Realfakes. Facta Simonidis. 2026;19(1):133-168. doi:10.56583/fs.2927
BibTeX
@article{Kabaa2026,
author = {Kabała, Mikołaj},
title = {Artificial Intelligence and Security in the Age of Realfakes},
journal = {Facta Simonidis},
year = {2026},
volume = {19},
number = {1},
pages = {133--168},
doi = {10.56583/fs.2927}
}
EUVIMED