For much of the 15 years covered by this publication, the human discourse on the ethics of artificial intelligence (AI) was characterized by what may be called “ambivalence” rather than concern. At least from the initial breakthroughs in neural network learning starting in 2006 to the deep learning and convolutional neural network enhancements starting soon after, AI has increasingly been seen to promise, if not yet to deliver, an impressive array of capabilities. But the negatives began to surface and became more commonly discussed around the time of the publications in 2017, as exemplified by the principles and the ethics report on the Governance of AI, particularly the less inclined principles about superintelligent algorithmic and networked information technology-based generalized optimization systems and the military and commercial control of information technology entities with bank balances and business contracts rather than hearts and souls. There, at least some people preparing for the future world are aware that developing systems that can resist external malice while continuing to provide general self-improvement, and systems that can retain and defend the ethical and legal definitions and directives concerning human welfare, dignity, and rights, is not like putting a man on the moon. Ethical concepts such as autonomy, vulnerability, deception, neglect, and so forth have strong roots in human-based normative reasoning and resultant actions. These ten concepts provide the basis for a design framework considering the deliberation and actuation enabling required for achieving ethical AI system behavior. Care to avoid, prevent, or provide remedy for violations of one or more of these ten rudimentary ethics concepts during an AI system life cycle are essential for real-world technical systems, meaning those used beyond today’s small-scale conversational prototypes, if AI technologies are to be fair, stable, reliable, controllable, and safe societal assets.