The article discusses a study conducted by researchers at the University of Oxford and the University of Naples Federico II, where they found that people generally find AI-generated food photos more appealing than real ones. The study used DALL-E 3 integrated with ChatGPT 4 to create images of natural, processed, and ultra-processed foods, which were then rated by participants on their appetizing qualities.
The researchers discovered that participants preferred AI-generated food photos, especially on the ultra-processed end of the spectrum, when they didn't know how the image was created. However, when the nature of the pictures was disclosed, participants showed a preference for real food images. The study highlighted that features such as glossiness, lighting, shape, and symmetry played a role in making AI-generated images more appealing.
The article also points out that using AI-generated food images could potentially lead to unrealistic expectations of how food should look, potentially impacting efforts to reduce food waste. Additionally, the study found that the AI-generated images tended to showcase higher-calorie foods in greater abundance, potentially influencing unhealthy eating behaviors among consumers.
Overall, the study suggests that while AI-generated visuals may offer cost-saving opportunities for the food industry, there are potential risks associated with exacerbating 'visual hunger' among consumers, which could lead to unhealthy eating habits or unrealistic expectations about food.
Comments