The members of a NATO Information Systems Technology (IST) Panel Exploratory Team (ET) spent the last year examining issues related to Machine Learning (ML) techniques for application in training and simulation. ML is an enabling technology that focuses on adapting computer behaviour on the basis of examples. In general, ML techniques aim at improving a computer program's performance of a certain task through experience. The group mainly focused on the application of ML in the development of behaviour of intelligent agents. Currently automated intelligent behaviour of entities in training simulations still relies heavily on the use of predefined scripts. Prerequisite for successfully creating autonomous intelligent agents, to be used in training, are the required underlying Artificial Intelligence (AI) techniques. A focus on ML techniques provides the means to establish autonomous intelligent behaviour of battlefield agents that goes beyond scripted behaviour. This paper presents the group's evaluation of ML techniques for training and simulation and the impact of these techniques on the development of intelligent agents in simulation applications.