A Note on This Book's Name

The vocabulary around understanding and governing AI can feel overwhelming. You will encounter terms such as responsible AI, interpretable machine learning, explainable AI (XAI), transparent AI, fair AI, and trustworthy AI. These terms overlap, but they are not exact synonyms. Each emphasizes a different question: Can people understand the model? Can they explain a particular decision? Is the system documented? Does it treat people fairly? Is it reliable, accountable, and appropriate to use?

This book is called Explainable AI for a few reasons:

• It signals a technical focus. Terms such as responsible AI and trustworthy AI often emphasize sociotechnical questions involving bias, fairness, governance, and the effects of AI on people and institutions. Those questions are important and appear throughout this book, but our main focus is technical: understanding, evaluating, and explaining how models behave. XAI is also a recognizable industry term that appears in technical research, tools, teams, and job postings more often than many of the alternatives.
• It fits the breadth of the course. We cover many topics rather than treating one method in depth. Most of those topics still fit comfortably beneath the broad XAI umbrella.
• It is a cool acronym. Naming a book is serious business, but not so serious that we cannot appreciate a good acronym.

No single label captures the entire field, and an explanation alone does not make an AI system fair, safe, responsible, or trustworthy. We use Explainable AI as the clearest name for the technical center of this book, not as a replacement for the larger sociotechnical work surrounding AI.