Data security when using artificial intelligence systems in criminal proceedings

The relevance of this research is determined by the dualistic nature of the implementation of artificial intelligence systems in criminal proceedings. On the one hand, automated collection and analysis of big data creates new risks for information privacy and individual rights. On the other hand, artificial intelligence technologies themselves possess significant potential for enhancing data protection through predictive threat analysis, access control, and transparency. This paradox requires careful scientific understanding. The subject of this research is the legal, organizational, and technical mechanisms for ensuring data security when using artificial intelligence systems in criminal proceedings. The aim of the study is to demonstrate, through a comparative analysis of international experience (Germany and China), that appropriate legal regulation can transform the use of artificial intelligence systems from a source of threats into an effective tool for ensuring data security in criminal proceedings. The article employs a combination of general scientific and specialized methods. The basis for this research is a comparative legal analysis of criminal procedural regulation and AI application practices in Germany and China. The formal legal method was also used to analyze regulatory acts, and the system analysis method was used to study the architecture of AI systems as an element of procedural safeguards. The scientific novelty lies in the substantiation of the thesis that data security is ensured not by prohibiting or limiting the use of modern technologies, but by their procedural integration. Using the German model as an example, it is demonstrated that consistent legal regulation (the principle of targeted restriction, access gradation, and encryption requirements) creates the basis for the legitimate use of specialized artificial intelligence systems. The Chinese model of legal regulation also ensures data security, but not through strict prohibitions and restrictions, but through the centralized architecture of the AI system used in criminal proceedings. It is concluded that many risks associated with the use of AI systems are transitional in nature and can be overcome through legal regulation and technological development. Proposals have been formulated to improve current legislation: establishing the status of electronic data, “technologically neutral” encryption requirements, establishing criteria for admitting non-governmental developers, and obligations to ensure the interpretability of artificial intelligence solutions.