Detection of the presence of people indoors is imperative for security in both public institutions and households. This is typically achieved through the use of cameras, which possess several drawbacks. An alternative method for detecting people using a Wi-Fi field is proposed. This approach respects privacy, is cost-effective, particularly given that the majority of individuals already possess Wi-Fi transmitters. It leverages the fact that the human body exhibits a high relative permittivity at Wi-Fi frequencies. Thus, merely by standing in a Wi-Fi field, individuals inadvertently alter its properties, obviating the need for additional devices. The data gathered from an experiment measuring RSSI values of the Wi-Fi field for various group sizes in a room were subjected to machine learning. The aim is to develop a model capable of predicting the number of people in the room based on unseen RSSI values. The text provides an overview of the physics of electromagnetic waves, their interaction with matter, and the fundamentals of wave propagation in communication systems. It also outlines the theory of shallow learning algorithms, such as logistic regression, support vector machines, k nearest neighbors, decision trees, and the random forest algorithm. Furthermore, it offers an introduction to the basics of neural networks. The results indicate very high prediction accuracies for the number of people in the room using all the mentioned algorithms. Specifically, all shallow models achieved accuracies exceeding 96 %, while neural networks surpassed 98 %. The highest classification accuracy on unseen data reached 99.2 %.