생성형 AI 기반 미디어아트의 상호작용 유형 및 특성 분석

Abstract

(Background and Purpose) The rapid advancement of AI technology has led to an increasing application of generative AI in the field of art, changing how audiences interact with media art. Generative AI-based media art extends the boundaries of artistic expression, allowing audiences to engage with an artwork dynamically and in real-time. However, despite the widespread integration of AI into media art, systematic studies of the interaction between generative AI-based media art and its audiences remain limited. Therefore, this study examines and classifies interaction types in generative AI-based media art, focusing on the processing of audience interactions through text and voice inputs to identify the unique characteristics of each interaction type. (Method) Representative examples of generative AI-based media artwork with audience interactions were selected. The interaction types were divided into two primary categories—voice-based and text-based—according to studies that classified interaction types within conversational interfaces. The concept of media art interaction was based on previous research, and a framework was applied to trace the process from audience input, through AI transformation, to the final performance outcome. Through this structured approach, each case study was examined to understand the influence of voice and text inputs on the modes of interaction and overall audience experience within media art. (Results) The analysis revealed that voice-based interactions deliver an intuitive and immersive experience. The experience allows audiences to participate naturally and facilitating immediate engagement through real-time feedback, which deepens immersion. In voice-based generative interactions, the AI system rapidly generates responses to audience voice commands, reinforcing the immediacy and intuitive nature of the interaction mode. Furthermore, voice-based image-generating interactions transform audience input into visual representations, offering a novel artistic experience that merges auditory and visual elements in real-time. Text-based interactions provide more structured and logical interactions, as audiences receive precise and targeted feedback in response to specific commands. This form of interaction is particularly effective in cases that require succinct or precise inputs. Text-voice combination interactions further enrich the experience by producing voice-based responses to text inputs, enabling audiences to explore more layered engagement. Additionally, text-image interactions integrate linguistic input with visual components, allowing audiences to experience a unique synthesis of language and visual art. (Conclusions) This study examined each subcategory of voice and text inputs. It found that voice input enhances audience immersion by emphasizing immediacy and intuitiveness, whereas text input supports structured and logical interactions, making it suitable for precise command input. The findings highlight the importance of considering each interaction type’s characteristics when designing generative AI-based media art. Future research should explore the impacts of these interaction types on artistic experience and expand the possibilities of AI-art fusion.