“Nudging” is the term used in the IEEE standards work on Ethics for AI Design. An AI system that applies deep learning to manipulating human decisions, with detailed analysis of the targeted individual, is a disturbing potential that must affect our trust in both the systems and those that direct their applications.
