AI 시대 데이터 주권에 대한 주관성 연구

Abstract

This study explores users’ subjective perceptions of data sovereignty in the age of artificial intelligence using Q methodology. Based on a review of literature and policy documents, a Q sort was conducted with 35 Q statements and 26 participants. The analysis yielded four types namely rights and regulation centered type, rights and development parallel type, context pragmatic type, and consent and compensation oriented type. The types differed in the priority given to rights and control, the preconditions for accepting efficiency and accuracy, expectations of compensation, and preferences for risk and domain specific differentiation. Across types, participants favored coupling rights guarantees with individual responsibility and accepted limited relaxation when purpose, duration, and scope were clearly specified. They emphasized execution over declaration, underscoring the importance of consent management, record keeping and auditability, and practical deletability. The study suggests acceptance conditions such as transparency, auditability, tools for exercising rights, and compensation design for debates on a basic artificial intelligence act, cross border data transfers, and automated decision making. Limitations include a sample with relatively high exposure to artificial intelligence and a modest size, and future research should broaden demographics and pair Q methodology with in depth interviews.