| Title |
An Empirical Study on Hyperreality in Architectural Visualization and the Limits of Expert Perception in the Age of AI |
| DOI |
https://doi.org/10.5659/JAIK.2026.42.1.61 |
| Keywords |
hyperreality; AI-generated images; architectural visualization; expert perception; metacognition |
| Abstract |
This study examines the hyperreality produced by AI-generated imagery in architectural visualization. Thirteen architectural professionals with
at least fifteen years of experience evaluated forty-eight images, half created by AI and half captured as photographs. The overall accuracy
rate was 52.40 percent, a result statistically indistinguishable from chance at 50 percent, showing that experts cannot reliably distinguish
AI-generated images from real photographs. Authentic architectural photographs, especially exterior shots, which showed an accuracy rate of
42.31 percent, were misidentified as artificial more often than the AI-generated images themselves. This supports Baudrillard’s concept of
simulacra in architectural visualization, in which AI imagery becomes hyperreal and appears more convincing than reality, signaling a
displacement of the real by simulation. Metacognitive results showed notable distortions. Although participants reported high confidence,
averaging 3.87 out of 5, the relationship between confidence and accuracy was minimal. The Cognitive Dissonance Index showed that 92.31
percent of participants were overconfident, and 41.28 percent of their highest-confidence judgments were incorrect. These outcomes indicate
that experts lack accurate awareness of their evaluative abilities. Overall, the findings reflect a major shift in the era of AI-driven
architectural visualization. As AI imagery becomes indistinguishable from physical reality, traditional authenticity-based evaluation loses
relevance. Future assessment approaches must focus on communicative clarity and interpretive value. This study offers empirical evidence of
AI’s transformative influence on architectural visualization and highlights the urgent need for new critical methods and educational frameworks
suited to hyperreal environments. |