Automated Machine Learning is an approach for automating the machine learning process by automatically selection of the most suitable machine learning algorithm and tuning its hyperparameters to create a machine learning model. Use of AutoML methods for prediction of traffic accidents severity can help improve the quality of models used to estimate the probability of different accidents based on various factors such as weather conditions, road and vehicle types, and driver behavior. The use of AutoML can significantly reduce the time required to create and tune models, as well as improve the accuracy of traffic accidents severity predictions, which in turn can lead to more efficient traffic management and fewer accidents. In this work we explore the applicability of different Auto ML libraries to the task of traffic accidents prediction and compare them with manually selected and tuned algorithms.