The Meeting of Generative Artificial Intelligence (GAI) and the Large Language Model (LLM): An Application on Book Covers

Abstract

In parallel with the development of computer technologies, artificial intelligence and related technologies are also developing rapidly. Artificial intelligence is manifested with practical and helpful applications in almost every field. One of the most remarkable of these applications is creating visuals with generative artificial intelligence. Generative AI refers to artificial intelligence models designed to generate new content in the form of written text, audio, image or video. Applications such as Midjourney, Dall-E and Leonardo AI are known to be the most preferred applications for generating images with generative AI. What these applications have in common is that a text input is given for the image generation process. While text input alone may be sufficient, more detailed and desired results can be obtained by specifying some parameters.

This study redesigned the covers of X children's books using Gemini and Midjourney artificial intelligence tools. The children's book covers were first analyzed textually with Google's Gemini artificial intelligence tool, and the outputs obtained were entered into the Midjourney artificial intelligence image generation tool with specific parameters to obtain visuals. In this context, the study is designed in a way that has never been addressed before in the literature, and in this respect, it stands out with the fact that the human factor is almost never involved in the process. In the study, how Gemini and Midjourney, one of the most up-to-date technologies, are used is explained in detail for the benefit of future researchers, and the results obtained are analyzed by matching the original data. The study will significantly contribute to the field as it is the first research in this field and explains the technique with its parameters.

Keywords: book cover, graphic design, fine arts, artificial intelligence, gemini.

DOI: 10.7176/RHSS/14-5-10

Publication date: June 30^th 2024