인공지능 시대 한국 현대시의 생성시학

Abstract

This article aims to diagnose the effects and prospects of the incorporation of artificial intelligence (AI) in contemporary Korean poetry by articulating the characteristics of AI poetry through the concepts of “generative poetics” and “blurriness.” Starting from the premise that the blurriness of generative poetics is linked to the essence of poetry, I trace its poetic genealogy to Kim Chun-su’s “artificial poetics” and postmodern poetics in order to view generative poetics as an extension of the history of contemporary Korean poetry. By referring to the poems of a human poet who attempted to perform the poetic utterance of “becoming-AI,” the poems created in collaboration between a human poet and an AI poet, and the poems of an AI poet who learned the poems of human, I try to diagnose the possibility of generative poetics in contemporary Korean poetry in the age of AI. To do so, I examine the poetic (im)possibilities in patterned poetic devices, the poetic possibility of interpolation, the poetic subject’s desire for “becoming,” and human-machine ensembles. In “Rhetorical Devices of Deeply Learned Patterns,” I investigate 1 the principle of metaphorical language through the paradigmatic axis (of selection) and the syntagmatic axis (of combination), 2 poetic structuring through repetition, symmetry and concatenation, 3 paradox and irony through opposite or contradictory statements, 4 mathematical, scientific, and musical principles, as the major poetic expressions. “Poetic Ambiguity of Interpolation” analyzes the notion of “interpolation”, the modus operandi of AI, as a principle of poetic creation of connecting, superimposing, and correlating. Thus, the poetic possibility of generative poetics is explored in terms of ambiguity, which is one of the most essential characteristics of poetry. “Fictional Poetic Subject and the Persona of ‘Becomin’” deals with the poetic subject in generative poetics. I argue that the poetic subject in poetry generated by/with AI, like a poetic subject in contemporary poetry in general, utilizes a speaker who dreams of “becoming” through “I/(me)”. Through poems that present an AI poet as the speaker, I examine the possibility of AI’s poetic self-consciousness, emotions, or thoughts via the concept of the self and identity of an AI. In “Creative Human-Machine Ensembles and Networks,” I look for specific clues of human-AI co-evolution through prompt-based poetry creation, and I search for the possibility of transcendent solidarity and collective organization of human-machine ensembles through empathy and emotion. New poetry will transcend conventional poetic norms. AI is already changing the way humans experience, think, remember, and write poetry, and AI poets will continue to evolve to create new poetry in various relationships with AI. This co-evolution is the essence of generative poetics, which lies between human and machine, bridging a human (poet) and an AI (poet).