AI 기법을 활용한 감청의 법적 쟁점

With the development of modern information and communication technologies, it is undeniable that traditional methods of lawful interception have become virtually infeasible, and even packet interception has reached its limits. As one of the main entities responsible for executing interception, investigative and intelligence agencies now face the urgent need to examine how they can respond to these technological constraints and increasingly sophisticated forms of crime. Fortunately, the advancement of information and communication technologies has also brought about the development of AI technologies, ultimately leading to the emergence of AI-based interception. However, in Korea, this field remains largely undefined, and thus this paper seeks to establish a conceptual framework for AI-based lawful interception. Along with defining the concept, several foreseeable legal issues are identified; nonetheless, the study confirms that such interception techniques can be promptly implemented without requiring separate amendments to the Protection of Communications Secrets Act, the statute most directly related to lawful interception. To this end, the study first examines how interception methods have evolved in accordance with the development of communication infrastructures, comparing the characteristics of interception during each technological stage. The analysis shows that traditional interception methods, which relied on collecting analog signals, shifted to packet interception as all communications became digitized. Today, however, even packet-based interception has become increasingly impracticable due to the widespread adoption of encryption. Despite these challenging circumstances, AI-based interception is emerging internationally, prompting an evaluation of the necessity and effectiveness of AI technologies in actual interception contexts. The results demonstrate that AI-based lawful interception can overcome the limitations of traditional approaches. Nevertheless, although current AI technologies have indeed advanced in ways relevant to interception, concerns persist regarding their technical complexity and the opacity of their decision-making processes. Additionally, because the fundamental purpose of lawful interception is the protection of individuals’ communication privacy and the promotion of freedom of communication, there remain fears that AI-based methods may expand privacy intrusions. These legal concerns necessitate conceptual clarification before further debate can proceed. Accordingly, this paper defines AI-based lawful interception as “a technical means that may be auxiliary to the acquisition or recording of telecommunications, utilizing AI or other automated tools to collect or analyze large-scale data or traffic patterns.” Based on this definition, the paper presents the legal issues that may arise from its application.