Effectiveness of Generative AI Tool to Determine Fruit Quality: Watermelon Case Study

To select a good quality watermelon, one needs the ability and experience to
recognize specific patterns in its visual characteristics. As buyers usually cannot taste
the watermelon beforehand, the outer patterns of a good quality watermelon may vary
depending on the perspective of the purchaser. As a result, there is a gradual adoption
of new generative artificial intelligence (AI) tools in the field of horticulture. These tools
are expected to minimize bias in human perception when determining the quality of a
watermelon based on its outer characteristics. This study aimed to compare the quality of
watermelons selected by generative AI with a panel sensory evaluation test. The results
of the two case studies indicate a significant difference in the quality of the generative
AI-selected watermelons. As an average, watermelon evaluators favored the watermelons
selected by ChatGPT as the best based on the Wilcoxon rank sum test and paired t-test
(p < 0.05). In conclusion, watermelons can be selected by ChatGPT with minimal effort,
promptly meeting consumer expectations.