FORMATION AND IMPROVEMENT OF A CRYPTOGRAPHIC KEY MANAGEMENT POLICY BASED ON THE GROUNDED AI CONCEPT

Abstract

The article is devoted to investigating the problems of forming and improving cryptographic key management policies in modern information and communication systems based on the Grounded AI concept, which supports decision-making through the use of verified data, contextual information, and explainability mechanisms. It is shown that traditional key management policies based on static rules and fixed key rotation intervals do not provide the required level of adaptability under dynamically changing conditions of the information environment, user characteristics, resources, and cyber threats. The specific features of applying the Grounded AI concept to support the generation, distribution, storage, use, renewal, revocation, and destruction of cryptographic keys are considered with due regard to the current operational context of the system. The factors influencing the selection of a key management policy are investigated, including the level of risk, the criticality of information resources, trust in interacting entities, execution environment characteristics, and regulatory requirements. An approach to forming an adaptive cryptographic key management policy is proposed, in which decisions concerning the key life cycle are made on the basis of an integrated analysis of contextual parameters, risk assessment, and the outputs of Grounded AI models, with the possibility of providing justification for the decisions made. A conceptual model of interaction among policy components has been developed, combining modules for monitoring, context assessment, risk analysis, recommendation generation, decision explanation, and implementation of control actions. The concluding section substantiates the advantages of the proposed approach, which include increasing the adaptability, controllability, transparency, and validity of cryptographic key management processes, reducing the probability of key material compromise, and improving the operational efficiency of cryptographic information protection systems. According to the scenario-based evaluation results, decision-making accuracy increased from 91.8% to 98.4%, the average time required to generate a control action decreased from 214 to 132 ms, and overall efficiency increased from 89.6% to 98.1% compared with a static policy.