생성형 AI 기반 비주얼 노벨 저작 도구의 제작 효율 인식과 워크플로우 재설계

초록

Visual novel production is a labor-intensive, multistage process involving planning, art, scripting, and engine building, with tool switching and repetition serving as major bottlenecks. This study analyzed the perceived efficiency and reuse intention among 155 users of VN-AI Studio, an artificial intelligence (AI)-based authoring tool that integrates the full pipeline from natural language prompts to final engine builds. Paired t-test results showed a 97.7% average reduction in production time (t = 8.565, p < 0.001). Approximately 74.2% of participants reported needing less than 20% manual modification of AI outputs. Regression analysis indicated perceived productivity as the key predictor of reuse intention (β = 0.677, p < 0.001), and cognitive load had a significant negative impact (β = -0.291, p < 0.001). Meanwhile, game development experience level showed no significant effect (F = 0.276, p = 0.843). These findings demonstrate that AI-integrated tools can function as an AI–Native Pipeline, redesigning the creation process and expanding access for small teams and nonexperts.

키워드