다큐멘터리에서의 AI 사용이 던지는 문제와 가능성: IDFA 개막작 <About a Hero>(2024)를 중심으로 = Challenges and Possibilities Brought by the Use of AI in Documentaries: Focusing on the Opening Film of IDFA <About a Hero> (2024)

다국어 초록 (Multilingual Abstract)

Centered around the concept of AI, About a Hero (2024) challenges conventional documentary classification by incorporating fictional narratives into a documentary framework through AI technology. This paper examines whether the work can indeed be classified as a documentary and how the use of AI reshapes traditional documentary conventions. Drawing on Bill Nichols’ documentary theory, a comprehensive analysis from the perspectives of creators, texts, and audiences reveals that while the film’s text may superficially resemble fiction or mock documentary, it is difficult to categorize as such. This is due to the director’s intent, which is not parody or genre critique but showcasing an artistic experiment utilizing AI, as well as the audience’s perception of the work not as mere fiction but as a real-world experiment. Additionally, the inclusion of expert interviews within the film strengthens its position as a documentary.
Based on this analysis, this study interprets About a Hero as a reflexive documentary that critically examines the creative process enabled by AI technology. While traditional reflexive documentaries have primarily focused on reflecting on expressive techniques such as interviews, narration, and archival footage, this work expands the scope of reflexivity by reevaluating the authenticity and meaning of AI-generated images and sounds. In doing so, it prompts audiences to engage in critical thinking about their overall consumption of images.
Furthermore, this study highlights how About a Hero possesses a performative dimension in addition to its reflexive nature, analyzing shifts in performativity caused by AI. The film-within-a-film created by AI Kaspar may seem foreign to the documentary form, yet this study interprets it as an example of AI-driven performativity. Specifically, the AI-generated Herzogian storytelling and deepfakeimplemented Herzog character exemplify an expansion of performativity, with the machine itself becoming an agent. This extension of performative modes to non-human subjects allows for diverse and novel storytelling possibilities.
Rather than dismissing AI-generated illogical results or unexpected imagination as mere technical errors, it is necessary to recognize them as distinct aesthetic characteristics of AI creation. While this phenomenon may be transitional, ongoing critical research should explore new outcomes presented by machines beyond human-centered creativity and investigate the collaborative possibilities between AI and humans