A con man deception can happen in service oriented architectures. The victim of this deception is the consumer. Hence, an algorithm dealing with such deception must to be consumer centric. It is desirable for such algorithms to learn the consumer perspective of the deception. In this work we incorporated a learning ability to the con man resistant trust algorithm. This is accomplished by empirical work with a machine learning algorithm for the con man resistant trust model that learns about a particular consumer perspective using the consumer’s historical data from the service. The results support a set of principles that the con man resistant trust algorithm follows, and a machine learning algorithm provides the con man resistant trust algorithm parameter setting for the specific consumer.