인공지능 시대, 한국 현대시 연구의 과제와 전망—생성형 인공지능 챗GPT를 중심으로

Abstract

The purpose of this paper is to review the challenges faced by Korean modern poetry research and to seek its direction at this point in time when we are witnessing the arrival of the era of artificial intelligence in earnest. As a premise for this, the results produced by generative artificial intelligence were reconsidered as a text concept as a phenomenon. The concept of text as an intra-action phenomenon in a process of intertwining without definitively dividing the subject and the object helps to identify the relationship between humans and ChatGPT. This is because the output produced by generative artificial intelligence is formed in the exchange of quantum. The big data accumulated by mankind is acquired by ChatGPT, the human’s primary question (command/request) is raised, the ChatGPT that collects the primary question gives the primary response (execution), and the human secondary question after reading the generated text Text is produced in the form of a feedback structure in which (command/request) follows, and the secondary response (execution) of ChatGPT, which has converged secondary questions, is fed back. Next, this paper paid attention to the analysis of poetry using ChatGPT. This paper divides this into two, one is the task of introducing research using ChatGPT for poetry before the text produced by artificial intelligence comes out, and the other is for poetry after the text produced by artificial intelligence comes out. As a result, the task of exploring a new research methodology was set. In the case of the task of conducting research using ChatGPT for poems before the text calculated by artificial intelligence came out, there were not a few cases in which inaccurate information was output due to hallucination. Therefore, this paper judged that the probability of deriving meaningful research results using generative AI at this point is low. In the case of the task of exploring a new research methodology for poetry after the text produced by artificial intelligence, it was argued that it was impossible to apply the existing author (poet) theory and work theory as it is. This is because text as a phenomenon produced by artificial intelligence is grasped only from a human-centered point of view, resulting in errors in analysis. Therefore, this paper emphasized the need to expand research perspectives that encompass humanistic criticism and machine criticism, deepen cultural theoretical poetry studies, and shift the category of poetics limited to literature to the activity of writing as a task for research on contemporary Korean poetry in the era of artificial intelligence.