Research Article

Constructing the algorithmic gaze: viewer positioning in AI-generated cinematic images

Volume: 8 Number: 15 October 15, 2026
TR EN

Constructing the algorithmic gaze: viewer positioning in AI-generated cinematic images

Abstract

This study examines how cinematic gaze is reconstructed within generative artificial intelligence systems. It argues that AI-supported image generation processes do not merely produce visual content, but also establish algorithmic regimes of spectatorship that organize the viewer’s position within the scene. To investigate this process, the same cinematic scene was regenerated through differently structured prompts using Midjourney V6, Leonardo AI, and Stable Diffusion based systems. The research combines a practice-based approach with comparative visual analysis. The generated images were analysed through three main categories: viewer-addressing gaze, viewer-aligned gaze, and external observer positioning. Particular attention was given to facial orientation, gaze direction, camera alignment, and spectator positioning. The findings demonstrate that generative AI systems tend to privilege frontal, highly readable, and viewer centred compositions. The study also reveals that different platforms respond to identical prompts through distinct visual organizations, suggesting that image generation is shaped not only by user input but also by model-specific aesthetic tendencies and learned visual biases. In several cases, the systems maintained dominant visual patterns even when prompts explicitly attempted to redirect spectator positioning, indicating forms of prompt resistance and algorithmic preference. The article argues that AI generated imagery should be understood not simply as aesthetic output, but as a post-cinematic visual regime that reorganizes perception, spectatorship, and visual orientation through algorithmic processes. By bringing together film theory, digital visual culture, and generative AI studies, the study aims to establish a theoretical bridge between cinematic gaze and algorithmic visuality..

Keywords

Ethical Statement

This study does not require an ethics committee approval

References

  1. Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (pp. 610–623). Association for Computing Machinery. https://doi.org/10.1145/3442188.3445922
  2. Bianchi, F., Kalluri, P., Durmus, E., Ladhak, F., Cheng, M., Nozza, D., Hashimoto, T., Jurafsky, D., Zou, J., & Caliskan, A. (2023). Easily accessible text-to-image generation amplifies demographic stereotypes at large scale. In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency (pp. 1493–1504). Association for Computing Machinery. https://doi.org/10.1145/3593013.3594095
  3. Bolter, J. D., & Grusin, R. (1999). Remediation: Understanding new media. MIT Press.
  4. Bordwell, D., & Thompson, K. (2013). Film art: An introduction (10th ed.). McGraw-Hill.
  5. Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20(1), 37–46. https://doi.org/10.1177/001316446002000104
  6. Elsaesser, T., & Hagener, M. (2010). Film theory: An introduction through the senses. Routledge.
  7. Hansen, M. B. N. (2016). Algorithmic sensibility: Reflections on the post-perceptual image. In S. Denson & J. Leyda (Eds.), Post-cinema: Theorizing 21st-century film (pp. 785–813). Reframe Books. https://doi.org/10.25969/mediarep/13477
  8. Leivada, E., Murphy, E., & Marcus, G. (2023). DALL·E 2 fails to reliably capture common syntactic processes. Social Sciences & Humanities Open, 8(1), Article 100648. https://doi.org/10.1016/j.ssaho.2023.100648

Details

Primary Language

English

Subjects

Cultural Studies (Other), Art History, Theory and Criticism (Other)

Journal Section

Research Article

Publication Date

October 15, 2026

Submission Date

May 25, 2026

Acceptance Date

August 25, 2026

Published in Issue

Year 2026 Volume: 8 Number: 15

APA
Aktaş, S. (2026). Constructing the algorithmic gaze: viewer positioning in AI-generated cinematic images. Uluslararası Sosyal Bilimler Ve Eğitim Dergisi, 8(15), 517-544. https://doi.org/10.67567/usbed.1958517
AMA
1.Aktaş S. Constructing the algorithmic gaze: viewer positioning in AI-generated cinematic images. USBED. 2026;8(15):517-544. doi:10.67567/usbed.1958517
Chicago
Aktaş, Seda. 2026. “Constructing the Algorithmic Gaze: Viewer Positioning in AI-Generated Cinematic Images”. Uluslararası Sosyal Bilimler Ve Eğitim Dergisi 8 (15): 517-44. https://doi.org/10.67567/usbed.1958517.
EndNote
Aktaş S (October 1, 2026) Constructing the algorithmic gaze: viewer positioning in AI-generated cinematic images. Uluslararası Sosyal Bilimler ve Eğitim Dergisi 8 15 517–544.
IEEE
[1]S. Aktaş, “Constructing the algorithmic gaze: viewer positioning in AI-generated cinematic images”, USBED, vol. 8, no. 15, pp. 517–544, Oct. 2026, doi: 10.67567/usbed.1958517.
ISNAD
Aktaş, Seda. “Constructing the Algorithmic Gaze: Viewer Positioning in AI-Generated Cinematic Images”. Uluslararası Sosyal Bilimler ve Eğitim Dergisi 8/15 (October 1, 2026): 517-544. https://doi.org/10.67567/usbed.1958517.
JAMA
1.Aktaş S. Constructing the algorithmic gaze: viewer positioning in AI-generated cinematic images. USBED. 2026;8:517–544.
MLA
Aktaş, Seda. “Constructing the Algorithmic Gaze: Viewer Positioning in AI-Generated Cinematic Images”. Uluslararası Sosyal Bilimler Ve Eğitim Dergisi, vol. 8, no. 15, Oct. 2026, pp. 517-44, doi:10.67567/usbed.1958517.
Vancouver
1.Seda Aktaş. Constructing the algorithmic gaze: viewer positioning in AI-generated cinematic images. USBED. 2026 Oct. 1;8(15):517-44. doi:10.67567/usbed.1958517

www.dergipark.org.tr/usbed

Editor in Chief:  Prof. Dr. Aytekin DEMİRCİOĞLU