Exploring the Artistic Stance of AI-Generated Images

Date

2025-1-31

Author

Kadıoğlu, Dilek

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

34
views

0
downloads

Cite This

An important question in the context of AI-generated images is whether they are art. The recent generative-AI systems generate images by means of text-to-image generation which are, in many cases, indistinguishable from human-made images. This indistinguishability has led to situations in which AI-generated images were mistaken for human art. Once the truth is revealed, however, these images lose their artistic stance. This dissertation argues that if some human-made images can have a claim to be art, then some AI-generated images should be allowed to have a similar claim. The indistinguishability in case does not merely stem from a resemblance due to copying human content. An examination of the architecture of the recent generative- AI systems reveals that the operations of these systems are much more complex than copying human content. AI-generated images exhibit some fundamental properties that are associated with human art which might be responsible for the indistinguishability in case. And such properties might also grant AI-generated images an artistic status. In this dissertation, I propose an output-oriented approach which is in conformity with human interaction with art objects and which solves many problems related to the evaluation of AI-generated images in terms of their artistic stance. My approach explains why the two sorts of images are hard to distinguish by employing the concept of creativity. It further proposes a modified version of the Turing Test as a suitable method for pursuing an output-oriented approach to AIgenerated images, namely, the Democratized Turing Test.

Subject Keywords

Generative-AI, The Turing Test, Creativity, Text-to-Image Generation

URI

https://hdl.handle.net/11511/113496

Collections

Graduate School of Social Sciences, Thesis