This chapter explores the relationship between the readability of textual prompts and the aesthetic qualities of AI-generated images. While prompt engineering has gained traction in design and creative fields, the specific role of linguistic structure, particularly readability, remains underexplored. Using data from an AI Drawing Competition held at the University of Liverpool in January 2025, the study analyses 20 prompts and their corresponding images generated using Midjourney. Five established readability indices were applied to assess textual complexity, while visual outputs were analysed using DenseNet-201, a convolutional neural network pretrained on ImageNet. Aesthetic metrics were extracted from the final feature layer of the network. A combination of qualitative analysis and a random forest regression model was used to explore whether readability metrics could predict visual characteristics. The model achieved strong performance, suggesting that prompt complexity significantly influences image aesthetics. Beyond quantitative results, the chapter offers a qualitative interpretation of how different linguistic structures elicit varying visual responses from generative AI. These findings highlight readability as a valuable parameter in prompt engineering, offering new insights into the design of human-AI creative interactions.