Investigating the Cognitive Differences in Scene Processing and Aesthetic Evaluation When Viewing Human- vs. AI-generated Imagery

The aesthetic experience of scene perception can be influenced by a variety of cognitive, perceptual, and oculomotor factors. Using a passive viewing eye-tracking paradigm, we explored the differences in visuospatial attention for free vs. guided viewing of human- vs. AI-generated scene photographs and paintings. Aesthetic ratings and fixation data were collected for each image, and heatmaps were generated to determine the spatial distribution of gaze. This work has the potential to reveal crucial differences in the visual processing and subsequent aesthetic evaluation of scene images of various sources and styles, and characterize what factors – if any – differentiate AI images from human creations.