Scientists at Siberian Federal University have developed a method for applying artificial intelligence in monitoring possible hazardous oil sites in the Arctic. Scientists decided to use neural networks to model a man-made disaster. Artificial intelligence was trained with the help of daily operational data on 14 main vectors of features that affect the rate of spread of the accident. New technologies of this kind are very much needed in the Arctic. So far, monitoring the condition of fuel tanks is problematic due to remoteness and difficult weather conditions. The existing methods of assessing the area of oil product spills have limitations. Therefore, the development of Siberian scientists will be in demand.