Safety culture (SC) is a modern paradigm of safety management of ultra-complex technical systems, primarily NPPs. According to the latest researches, the probability of an error made by a human, who professes the principles of safety culture, and the occurrence of accidents in general, compared to previ-ous methods of safety management, is reduced by hundreds of times. Therefore, it is necessary to pro-vide support as much as possible at all levels of government, to train all specialists and explore new pos-sibilities of this paradigm. This article is aimed at developing offers for an algorithm for automated as-sessment of the level of SC power units based on current indicators and indicators of structural units with a selected discreteness in order to further develop an appropriate cloud-based software which will eliminate subjectivism and record the smallest deviations, which, in its turn, will allow to take appropri-ate measures to improve the safety of nuclear power plants in the early stages of negative processes. The term «safety culture» was first formulated by the IAEA in 1986 in the course of the analysis of the causes of the Chernobyl disaster, and published in INSAG-1. The report stated that the lack of safety culture was one of the main causes of the disaster. Safety culture is defined as a set of values and ac-tions that occur as a result of an agreement reached in the team between the top managers and employ-ees to give security the highest priority in order to protect people and the environment. Further under-standing of this concept has led to a new perspective on the causes of other accidents and incidents at nuclear power plants that occurred in the past. The experience of NPP operation shows that their occur-rence is somehow related to people's behavior (namely, their attitude to safety issues), i.e., the current state of the level of safety culture. But as a current safety parameter (SC) it is estimated with a long pe-riod of time. The article offers automation of the assessment process.