PROSPECTS OF APPLICATION OF ARTIFICIAL INTELLIGENCE IN PLANT BREEDING

Increasing the efficiency of agricultural activity implies a transition to precision farming, the basis of which is digital technology. Selection work is characterized by high labour intensity and time costs in the development of new crop varieties. The solution to this problem can be the introduction of artificial intelligence in selection work. In this article, the main trends in the introduction of artificial intelligence in plant selection are presented by analysing the available practical research on the application of neural network models in the field of selection work. A number of positive consequences of such digitalization were found: increasing the accuracy of classification, prediction, efficiency and purposefulness of work; modelling the process of crossing and assessing the value of phenotypic properties; considering non-linear relationships in the formation of new species and adaptation to a particular habitat; collection and analysis of a large amount of data needed in selection process, etc. The article also shows the limitations and dangers of AI application: the complexity of training models for adequate work with data; the need for specialists combining training in both selection process and neural networks; the complexity of hypothesis testing; financial and energy costs of large models; collapse of generative models and cognitive distortions of neural network thinking; neuroethical problems of interaction of artificial systems with man and nature. The necessity of further development and improvement in the application of neural network models in solving selection problems is shown.