글쓰기 체계로서의 인공지능

초록

This study problematizes the prevailing view that evaluates Artificial Intelligence (AI) text through the lens of representation and demands technical refinements to improve its accuracy. This representationist-instrumentalist perspective frames AI merely as a tool for human ends, failing to grasp the fundamental shift in the horizon of meaning production brought about by AI. Accordingly, this study conceptualizes AI not as a tool for reality representation, but as a ‘writing system’ that generates meaning based on its own grammar. Employing Jacques Derrida’s deconstruction as a theoretical lens, the study argues that AI writing, much like human writing, operates under écriture— lacking an origin, author, or destination. Furthermore, the study critically examines the contradiction wherein AI, while operating solely within the signifying conditions ofécriture, simultaneously harbors alogocentric desire. Outputs from AI writing are frequently equated with objective truth and justified grounded in scientism. The study proposes that the ethical imperative of the AI era is not to secure technical accuracy or recover anthropocentric meaning, but rather to create fissures in AI’s writing, which attempts to subsume difference and meaning under a single logic.

키워드