METHODOLOGICAL ASPECTS AND LIMITATIONS FOR ENSURING SAFETY IN TRANSPORT MANAGEMENT USING ARTIFICIAL INTELLIGENCE

Recently, there has been rapid technological progress in transportation, which is expressed in the field of telematics, the Internet of Things, the Internet of Vehicles, Big Data Analysis, etc. The use of information technology by drivers, including so-called smart devices, technological advances in sensor devices, autonomous vehicles, vessel telematics, cameras, etc., provides new potential for driver/operator behaviour monitoring, vehicle communication, surveillance, and incident detection in all modes of transport. The adoption of these technological advances is widespread, and the emanation of their use is the introduction of A.I. into the decision-making process. All this opens up the important topic of taking responsibility for decisions, detecting and eliminating errors in decisions and actions, and hazardous traffic management and safety processes. This article aims to examine and summarize these problems, especially when using A.I. in transport, from the point of view of ensuring safety. Methodological formulations and a general concept for a limit of admissibility in ensuring the necessary safety in transport are also proposed.