Designing a Maritime Cybersecurity Risk Intelligence Model with Generative AI and Real-Time Interviews

AI has disrupted all sectors, and the maritime industry is undergoing its own transformation. After reviewing 48 academic papers, the authors showed the need for a holistic and engaging cybersecurity risk analysis method and suggested a survey-based method called Cyber Risk Analysis Method for Maritime Transportation Systems (CRAMMTS). This study presents an AI-powered system for conducting risk analysis interviews with maritime entities, developed using LangChain, LangGraph and OpenAI’s GPT-4o in Python. The system uses external sources, such as Maritime Attack Database and CRAMMTS survey results, as a data source, evaluates each question/response pair in real-time, determining the necessity of follow-up questions and ensuring thorough risk coverage. The system stores the interview results, applies basic analysis, and generates recommendation statements for the entity. Retrieval augmented generation is employed to integrate current incidents and risk data into the analysis. Future expansion opportunities include automated analysis and the generation of comprehensive risk analysis documents with actionable recommendations. This approach not only aims to streamline and standardize the maritime risk analysis process but also reduce the cost barrier to cybersecurity resources using advanced language models.