The combined effect of entropy and complexity: Human analysis of AI painting recognition ability

The use of AI-generated artistic images has seen extensive applications in the field of packaging design. However, not all AI-generated images are readily accepted by users. This study investigates how the attributes of artistic images influence individuals' perceptions of whether an image was created by AI. The focus is not on users' ability to identify AI-generated images, but rather on their "trust" in the method by which an image was created. This serves as an empirical reference for designers to select AI-generated images and provides insights for the development of AI image generation technologies.The study classified 160 art images according to three attributes: complexity, sample entropy, and artistic style, and asked participants to rate their confidence in whether each image was generated by AI.Additionally, participants took a questionnaire assessing their level of art knowledge. The study analyzes the relationship between the attributes of artistic images and the level of trust participants place in the method of image creation. The findings hold practical significance for understanding how user trust in artistic images is constructed and enhance our comprehension of how image attributes affect human perception.