Walvius, Tess (2026) Improving Human AI Image Detection: the Role of Emotional Content, Visual Attention to Detail Skills and Inductive Learning. Bachelor thesis, Psychology.
|
Text
Thesis-final-version-Tess-Walvius---s5490146---1.pdf Download (17MB) | Preview |
A thesis is an aptitude test for students. The approval of the thesis is proof that the student has sufficient research and reporting skills to graduate but does not guarantee the quality of the research and the results of the research as such, and the thesis is therefore not necessarily suitable to be used as an academic source to refer to. If you would like to know more about the research discussed in this thesis and any publications based on it, to which you could refer, please contact the supervisor mentioned.
Abstract
Recent advances in text-to-image generation have raised concerns about the potential harmful effects of this new technology: scams, manipulation and misinformation. The present study aimed to investigate the factors that influence humans’ ability to successfully differentiate between AI and real images. The effect of inductive learning on improving human AI (and real) image detection, as well as the effects of emotional content and visual attention to detail ability were studied using an online image detection task. Based on the literature it was expected that inductive learning, high visual attention to detail skills and fearful pictures (as opposed to happy and angry images) would have a positive effect on the detection of AI generated images. The data of N = 270 participants was analysed. The results showed training through inductive learning to be effective, especially when the image contained a happy expression. AI Images with fearful facial expressions were detected more accurately than those containing happy or angry expressions, only when individuals did not receive any training. Visual skills measured by the L-EFT did not show any effect on the detection rate of AI images, nor did it show an effect on the detection of real images. No interaction effects of training were found between visual attention to detail for training and emotional content. These findings suggest that inductive learning can be a promising approach for improving human AI detection and may help with minimizing the risks associated with text-to-image generation.
| Item Type: | Thesis (Bachelor) |
|---|---|
| Supervisor name: | Gutzkow, B. |
| Degree programme: | Psychology |
| Differentiation route: | None [Bachelor Psychology] |
| Date Deposited: | 23 Feb 2026 11:37 |
| Last Modified: | 23 Feb 2026 11:37 |
| URI: | http://gmwpublic.studenttheses.ub.rug.nl/id/eprint/6254 |
Actions (login required)
 |
View Item |