Development of an Expert Chatbot for Digital Forensics Using RAG Model Implementation
Minsoo Kim, Do‐Youn Kim, Yunji Park, Doowon Jeong · 2024
Digital forensics is an important evidence collecting technique in criminal justice system. However, due to the invisibility of digital evidence, the results of analysis vary depending on the expertise of the investigator. Though there has been a trend to use commercially available chatbots to minimize the divergence of investigator's expertise, the effect was limited by unprofessional response and lack of standardized information sharing system. Therefore, this study proposed the data construction method for constructing expert chatbot based on web crawling of digital forensics related domains. Then, the collected data was used to build the new digital forensics knowledge DB. In addition, this research suggested the RAG based chatbot model that can address the problems of existing chatbots, such as generating mis-information and limitation in updating train data, and generate specialized answers based on the digital forensics knowledge DB. As a result, this research presented the possibility of establishing information sharing system of digital forensics and utilizing the specialized chatbot based on the system. Additionally, it provided the foundation of a multimodal chatbot for analyzing various forms of digital evidence.