‘과학기술학’ 관점에서의 AI Agent 사례 연구: ‘젠스파크’ 사례를 중심으로 = Case studies of AI agent from a Science and Technology Studies perspective -Focusing on the case of ‘Genspark’-

다국어 초록 (Multilingual Abstract)

By 2025, ‘AI Agent’ are emerging as a new technological trend. However, whenever new AI-related services are announced, the prevailing approach tends to be fragmented, focusing solely on the progressive or negative aspects of the technology. What is important, however, is not to speculate on ‘uncertain possibilities,’ but to track and observe how people actually adopt and utilize AI in their personal lives and in industrial settings. This is because AI technology emerged within the context of human-human relationships, and this relationship will continue in the future.
In this study, we focused on Bruno Latour's “Science and Technology Studies” perspective to understand AI not as an independent entity but as an entity that constitutes itself in relation to humans. Through this, we aimed to propose a theoretical framework that provides insight into the relationship between humans, human society, and AI. The research method involved analyzing previous domestic and international studies to examine the approaches of science and technology studies and Latour's key concepts, and then proposing a theoretical framework that provides insight into the network of “human-artificial intelligence-human society.” Subsequently, the applicability of the science and technology studies-based framework was explored using “Genspark,” a representative example of an artificial intelligence agent.
The results confirmed that, from a science and technology studies perspective, humans, AI, and human society are not independent entities but interconnected entities. Through the analysis of Genspark, it was revealed that users are gradually ‘delegating(Delegation)’ more authority to AI agents, suggesting a restructuring of the human-technology relationship. In particular, while discussions on the development and value of AI and the risks of human replacement are active whenever AI agent services emerge, mentions of the resource infrastructure required to operate such services are limited or treated as separate issues. This phenomenon, where the principles of AI services are hidden in a ‘Black Box,’ reveals the underlying issues of massive resource consumption and social infrastructure imbalances.
The theoretical framework proposed in this study is expected to serve as a methodological approach for gaining insights into AI technology and providing possibilities for addressing the societal changes that AI agents may bring about.