From Words to Icons. Advancements in Prompting Text-to-Image Models for Logo Design with Artificial Intelligence

This paper presents an extensive review and empirical analysis of text-to-image models, focusing on their application in logo design and visual generation. Beginning with the seminal work by Goodfellow et al. on Generative Adversarial Networks in 2014, the paper traces the evolution of text-to-image technology through subsequent developments such as Generative Adversarial Text-to-Image Synthesis and Creative Adversarial Networks. The study explores the effectiveness of prompt engineering in guiding AI models towards desired outputs, showcasing the potential of prompt-based texting in enhancing creativity and generating logo visual content. Furthermore, the paper presents empirical research through a targeted questionnaire distributed to design and artistic communities. The questionnaire evaluates the effectiveness of AI-driven prompting text-to-image models in logo design, providing insights into the generated logos’ perception and quality. The findings underscore the potential of AI as a valuable tool in logo design while highlighting areas for further refinement and development.