In article ideas of use of algorithms of artificial intelligence in pedagogics are considered. The methodology of existence and functioning so-called «Collective Pedagogical Mega System» is presented. The idea of input of such ephemeral formation is necessary only that is modeling and well to understand system of collective pedagogical intelligence, not directly to understand, as it works, what laws in it operate. On its basis it is possible to predict pedagogical processes and the phenomena, to reveal new laws.

Article purpose – to show application of algorithms of collective reason in student teaching for effective acceptance of didactic decisions. The matter is that in real educational process, except known set of pedagogical conditions, the unknown reasons and factors which are difficult for expecting and providing in advance operate also casual, beforehand. They arise spontaneously, and more often, have stochastic character. These single the reasons insignificantly influence results of the forecast of a choice of methods of training, but in aggregate their influence becomes so significant that they radically can change prognostic conclusions. The problem also consists in that in concrete conditions from set of the known (provided) and casual (unexpected) reasons to allocate those factors which would provide the highest efficiency and productivity of employment. For these purposes the most suitable algorithm in realization of selection of methods of training, is, the so-called ant algorithm which, on the one hand considers likelihood character of changing parameters of influence, and with another – steadily gives high productivity of the application.

On a concrete example process of realization of application of ant algorithm for revealing of the best hierarchy of pedagogical conditions (factors) which define an optimum choice of a method of training is shown.