Machine Learning Within a Con Resistant Trust Model