Opportunities and limitations of using agent-based artificial intelligence technologies in the Russian oil and gas industry

The potential of using agent-based artificial intelligence technologies in the Russian oil and gas complex has been analyzed, considering the industry infrastructure specifics and industrial safety requirements. The subject of the research is agent-based AI technologies as a tool for optimizing business processes of the Russian oil and gas companies. The purpose of the study is to identify the possibilities of using agent-based AI technologies to improve the efficiency of production and management processes in Russian oil and gas companies and identify key barriers to their implementation. The methodological framework includes an analysis of digital initiatives of Russian oil and gas companies and foreign experience in implementing agent-based AI technologies, interpretation of autonomy levels classifications (Feng et al., KPMG), as well as an assessment of business processes risk level from the standpoint of industrial safety. It has been established that the existing gap between the declared potential of agentbased technologies and real implementation practices is due to the high criticality of the errors consequences and the lack of regulated mechanisms for responsibility distribution between the operator and the system. It has been shown that it is the most pragmatic at the current stage to use a format of limited autonomy of agent systems, in which a person retains the right to surgical intervention. The practical significance of the study lies in substantiating the need to develop an industry-specific AI access system for critical operations, including an autonomy levels classification, requirements for verifiability of solutions, and control regulations to prevent emergencies. The results obtained can be used in forming corporate standards for digital transformation and government policy in the sphere of AI technologies in the oil and gas industry.