AI-합성생물학 융합의 법적 규제 프레임워크 연구: 적응형 혼합규제 모델을 중심으로

The convergence of AI and synthetic biology holds the potential to dramatically increase efficiency in high-tech convergence industrial production, such as in the treatment of intractable diseases; however, it also poses new risks that are difficult to regulate under existing frameworks. As AI-generated or processed Digital Sequence Information (DSI) spreads across borders, and multiple actions combine during the commercialization of DNA synthesis technology and the accumulation of information resources, "regulatory lag" occurs where intervention is not timely. Furthermore, the "Collingridge dilemma" arises, wherein the risks of advanced technology are unrecognized and neglected in early stages, making regulatory intervention impossible once development has significantly progressed. This study identifies regulatory voids in terms of "subject, actor, and governance" within the current legal system centered on the Transboundary Movement of Living Modified Organisms Act. It examines how problems arising from these voids and regulatory lags interact, leading to the fragmentation of legal liability and the disconnection of specialized risk assessment functions. As an alternative to resolve these issues, this study proposes an "Adaptive Hybrid Regulation Model" and presents its functional significance. First, regarding the regulatory subject, the model advances the timing of regulation to include Digital Sequence Information (DSI), extending beyond the stage of living organisms. Second, it applies behavioral standards—such as screening, mitigation of harmful functions, and maintenance of logs and data lineage—to key actors in high-risk areas (e.g., AI model developers, platform providers, and DNA synthesis companies). These are differentially imposed as minimum standards (hard law) based on risk levels, while detailed requirements subject to technological change are designed to be rapidly updated through modular standards (soft law). Third, a common audit infrastructure is introduced to ensure supervision under unified standards even if regulatory governance is fragmented. Simultaneously, adaptability and auditability are secured by linking continuous standard updates with enforcement status through a dual regulatory impact assessment. Furthermore, to institutionally implement this model, this study proposes a "1+3 Legislative Strategy" that strengthens the legal basis of the National Bio Committee and links it with relevant individual laws to function as a coordination hub. The proposed Framework Act on Bio-Economy and Security is established as the overarching governance basis, functionally linking the Transboundary Movement of Living Modified Organisms Act, the Framework Act on Promotion of Artificial Intelligence and Creation of Trust Foundation, and the Synthetic Biology Promotion Act. This Framework Act is designed not to replace enforcement regulations but to focus on "connecting norms," such as joint deliberation and data standards. Specific enforcement is distributed among competent ministries or specialized agencies to prevent over-regulation and overlapping authority. By approaching the regulatory paradigm shift in the era of convergent technology from the perspective of "risks associated with the linkage of information and matter," this study presents design principles and legislative strategies combining hard law, soft law, and audit infrastructure. Ultimately, it aims to contribute to building effective governance that supports safe yet innovative technological development.