Schmidt, Leonard (2024) Who Can Tell if It's AI? Using Inductive Learning to Distinguish AI Art and Traditional Art. Bachelor thesis, Psychology.
|
Text
Bachelor-Thesis.pdf Download (730kB) | Preview |
Abstract
Frequently, generative AI has seen a massive spike in attention and influence. People are more and more confronted with AI-generated content, especially artworks. This development sparked debate not only about the sophistication of AI and difficulty distinguishing its content, but also about whether an AI can really mimic human creativity. We intended to investigate this in order to see whether we could potentially train participants to get better at distinguishing between traditional and AI-generated artworks, and on the other hand whether participants higher in art knowledge would benefit more from the training. 35 participants in the experimental condition received an online inductive learning paradigm, and then completed a test where they had to indicate whether a presented picture was made with AI or not. Relative to the 47 participants in the no-training control condition, results confirm our first hypothesis, as they show that the experimental condition was significantly better at correctly identifying the source of the artwork during the test. However, contrary to our second hypothesis, art knowledge did not seem to have an influence on this relationship. Despite the apparent sophistication of AI, these results suggest that there still might be something in us that enables us to notice AI-generated content, given proper training. Next to some strengths, such as exploratory work in a highly relevant field, some limitations such as sample limitations and the online environment are also discussed. Future research should address these issues by conducting laboratory experiments and improving the sample by focusing on art experts.
| Item Type: | Thesis (Bachelor) |
|---|---|
| Supervisor name: | Gutzkow, B. |
| Degree programme: | Psychology |
| Differentiation route: | None [Bachelor Psychology] |
| Date Deposited: | 10 Jul 2024 08:07 |
| Last Modified: | 10 Jul 2024 08:07 |
| URI: | http://gmwpublic.studenttheses.ub.rug.nl/id/eprint/3786 |
Actions (login required)
 |
View Item |