The book develops as a formal conversation between an AI engineer, Peter, and a humanities researcher, Suman. Each presents short chapters by turn, responding to issues raised in the previous. It is written so that specialist terms and concepts in AI research are accessible to a humanities specialist. Meanwhile, those concepts within the humanities are made to be easily grasped by an AI engineer.The main thread of the argument is provided by the ostensible objective of setting up some ground rules to design a personal AI-Assistant, i.e. an AI system that could act as a personal friend, consultant and confidante for every individual and be integrated into our everyday life. These rules apply to four large areas of AI development: recognition and identification, communication, explanation, and civility - the 4 parts of the book are devoted to each.In discussing these areas, the book also gives an accessible account of the current state of AI research in each of the four realms, as well as adding nuance to the underpinning assumptions of relevant technologies, and reflecting on their social implications.The broader ideological objective of the book is captured by the phrase 'practical philosophy', which demonstrates a commitment to bringing AI research and development to cohere with some conception of a good life for all.