Csomor, Dávid (2024) Who Can Tell If It Is AI: Can Inductive Learning Improve AI-Art Literacy? Bachelor thesis, Psychology.
|
Text
DavidCsomorS4304837ThesisFinal1.pdf Download (675kB) | Preview |
Abstract
AI-generated art has been undergoing a rapid development in the past years both in terms of quality and efficiency. The aim of the current study was to investigate how good people are at distinguishing AI from non-AI art; and whether this ability can be improved by training. The method of choice for the training was based on the inductive learning paradigm, which prioritizes intuitive pattern recognition, rather than learning based on factual information. Thus, participants in the experimental group were instructed to observe artworks labelled as AI or non-AI; and, subsequently, they were asked to complete a test where they could label the artworks as AI or non-AI themselves. The results showed that participants performed at chance level in the absence of training; while participants in the experimental group performed only slightly, but statistically significantly, better. Additional variables were also explored as potential influences on test performance, such as confidence; AI/non-AI art knowledge; and the types of the artworks. The results suggest that, while AI art may be indistinguishable from non-AI art at first glance, people may be able to develop a better intuition for AI art recognition. The study was limited in statistical power though, and the homogeneity of the sample also warrants more studies to establish better generalizability.
| Item Type: | Thesis (Bachelor) |
|---|---|
| Supervisor name: | Gutzkow, B. |
| Degree programme: | Psychology |
| Differentiation route: | None [Bachelor Psychology] |
| Date Deposited: | 30 Jul 2024 07:38 |
| Last Modified: | 30 Jul 2024 07:38 |
| URI: | http://gmwpublic.studenttheses.ub.rug.nl/id/eprint/4187 |
Actions (login required)
 |
View Item |