Landzettel, Timo (2024) Can We Unlearn to Get Tricked? An Inductive Learning Approach to Distinguish Between AI Paintings and Traditional Paintings. Bachelor thesis, Psychology.
|
Text
Final-Thesis.pdf Download (376kB) | Preview |
Abstract
Without training, people fail to reliably distinguish between AI artworks and human-made artworks. This experiment aims at investigating whether it is possible to teach people to distinguish between AI artwork and human-made artwork by using a spaced inductive learning paradigm. The first hypothesis states that an inductive learning approach, combined with spacing, can teach people to better distinguish between AI artworks and human-made artworks. The second hypothesis states that people who are more interested in arts benefit more from the training compared to participants who are less interested in arts. To investigate the hypotheses, an experiment was designed and conducted on Qualtrics. The final data includes responses of 82 participants, most of them being University of Groningen first-year Psychology students. Overall, participants who received training were 57,3% accurate in identifying the creation process of a picture. An ANCOVA yielded a significant result for the first hypothesis (p = 0.004; ηp2 = 0.102). Therefore, evidence was found that a spaced inductive learning paradigm can teach people to better distinguish between AI artworks and human-made artworks. There was no evidence found about an interaction effect between art interest and training effectiveness (p = 0.409). Thus, people high in art interest do not seem to benefit more from the training. Even with training, participants’ ability to distinguish is still close to chance level in our sample. Therefore, an improved training might be necessary to reliably teach people to detect a difference. However, more developed AI artwork creating software might counter improvements in response accuracy in the future. Keywords: Inductive Learning, AI art, Creative AI, Art recognition
| Item Type: | Thesis (Bachelor) |
|---|---|
| Supervisor name: | Gutzkow, B. |
| Degree programme: | Psychology |
| Differentiation route: | None [Bachelor Psychology] |
| Date Deposited: | 12 Jul 2024 08:26 |
| Last Modified: | 12 Jul 2024 08:26 |
| URI: | http://gmwpublic.studenttheses.ub.rug.nl/id/eprint/3840 |
Actions (login required)
 |
View Item |