Steenbergen, Dennis (2025) Who Can Tell if it’s AI? Applying Inductive Learning to the Detection of AI Photography. Bachelor thesis, Psychology.
|
Text
s5186951BachelorThesis.pdf Download (704kB) | Preview |
Abstract
Images generated by artificial intelligence models have very recently begun dominating the digital landscape, represented both in encountered content and discussion. These images are becoming increasingly sophisticated and high-quality, especially supposed photographs, being ever harder to distinguish from actual photographs. Given the dangers that artificially generated photography can potentially pose, the present study investigated whether people can be trained to become better at distinguishing genuine photographs from artificially generated ones, and whether motivational predisposition would play a role in this. In an online study, N = 194 participants were presented with a series of photographs, and had to quickly indicate whether it was AI-generated or not. Of these, 91 were subjected to an inductive learning paradigm prior to testing. Results showed that experimental performance did not significantly differ between the groups and could not be predicted with motivational predisposition, indicating that people can generally not detect well-generated images, even after training. However, the experimental condition was revealed to have a significant effect on response style, such that participants were much more likely to appraise a photograph as artificially generated, regardless of veracity. The supposed inherently deceptive nature of artificially generated photographs might induce heightened defensiveness in image appraisal.
| Item Type: | Thesis (Bachelor) |
|---|---|
| Supervisor name: | Gutzkow, B. |
| Degree programme: | Psychology |
| Differentiation route: | None [Bachelor Psychology] |
| Date Deposited: | 10 Jul 2025 07:25 |
| Last Modified: | 10 Jul 2025 07:25 |
| URI: | http://gmwpublic.studenttheses.ub.rug.nl/id/eprint/5419 |
Actions (login required)
 |
View Item |