생성형 AI 프롬프트 구성 체계 도출을 위한 이미지 생성 AI 활용 사례 연구 - 이미지 생성형 AI 프롬프트 구성요소 간 비교분석을 중심으로 -

Abstract

This study aims to understand the attributes of generative AI as a new design tool and to propose a basic structure for AI prompts, which are conversational commands in order to do their effective use. First, to examine the changes brought about by AI technology in the industry and the resulting shifts in AI-powered design field, we reviewed previous studies and analyzed media content related to AI from various mass media sources. And we aimed to define the trendy attributes of generative AI by exploring its utilization cases in domestic companies, as well as through previous studies and online workshops. Additionally, the fundamental structure of AI prompts was derived through a comparative analysis of exemplary prompts created by AI artist and award-winning works from AI competitions. The four attributes of generative AI that affect image creation—similarity, interactivity, dependency, and ethics—were derived and explained in detail. Based on this, AI prompts were categorized into three stages: design target (image content), shape characteristics (style), and specific conditions (condition). After defining the content and characteristics of the components within each category in detail, the basic composition system was structured and derived based on the categories, types, and content/characteristics of the prompts. The proposal of a basic framework for creating generative AI prompts, based on an understanding of the attributes of generative AI, is significant as a result of this study. It is hoped that future research will continue to actively develop generative AI into a more robust design tool.