The question of energy supply continuity is essential from the perspective of the functioning of society and the economy today. The study describes modern methods of forecasting emergency situations using Artificial Intelligence (AI) tools, especially neural networks. It examines the structure of a properly functioning model in the areas of input data selection, network topology and learning algorithms, analyzes the functioning of an energy market built on the basis of a reserve market, and discusses the possibilities of economic optimization of such a model, including the question of safety.