Stephen Cave, Kanta Dihal, and Sarah Dillon (eds.), AI Narratives: A History of Imaginative Thinking about Intelligent Machines. Oxford, Oxford University Press, 2020, 14,4 × 22,3 cm, 448 pp., ISBN 978-0-1988-4666-6

I begin with an old story from ancient Egypt, about an “ushabti” (a servant statue, made from wood or clay, but magically animated to do labour for souls in the ancient Egyptian afterlife). Commanded to carry out a simple repetitive task (in this case, drawing water), the ushabti in the story does its duty too well, flooding the home of its master. This tale, thousands of years old, reminds us that people have always imagined artificial ways to make life easier (whether magical wooden servants in ancient Egypt, or assembly line machines in our time). Stories like this show how human beings have always used fiction to explore such ideas, long before it was possible to make them reality (and today’s speculative fiction/sci-fi does the same job). Stories such as this one also suggest that artificial solutions to problems often result in new, unexpected problems; AI Narratives does not reference this story of the ushabti, but it carefully considers all the ideas just mentioned. In considering this conversation, AI Narratives looks at speculative fiction, from ancient Greece (Homer’s Iliad) to twentieth century cyberpunk sci-fi (William Gibson’s 1984 novel Neuromancer, and later stories inspired by it), in order to explore how ideas of technology have evolved, and how concepts such as “artificial” have changed. The book also covers a wide variety of global perspectives, in addition to the voices it considers from across time: humanists, feminists, political theorists, and African American perspectives are among the voices describing (and pointing out the missing narratives) that could (and should) be listened to by those developing AI technology today. One important idea, conveyed throughout the “narrative of narratives” presented by the book, is how our vision of what AI is shaped as much by science as it is by culture; humankind has entered into a kind of dialogue with itself (and its artificial extensions), as our expectations, dreams, hopes, and fears shape our creations, leading us to reevaluate our thoughts and ideas once again. This dialogue has been unfolding since the beginning of recorded history, in the first literature penned by human beings, and continues until today. As each era of human progress is considered, three pivotal questions were raised for me in each chapter of the book, with a different answer provided by each generation which tells stories and builds technology (thus offering an interesting insight into human progress, the evolution of ideas, and our own self-understanding). These questions, as they occurred to me, were: First, who creates? Second, why do creators create? Third, through whose authority do creators create? A fourth question emerges in the later chapters of the book, as AI is depicted mimicking human behavior more and more accurately): What (or who) is being created, and why? These questions often overlap; for example, the identity of the creator and the reason for the creation are linked, and there are times when the creation holds up a mirror to the creator’s face, allowing us to see who the creator really is, by virtue of its function, purpose, and the need it was created to fulfill. Consider the examples provided in the earliest chapters, covering classical Greece (specifically, the writings of Homer) and medieval European legends – the creators of these early stories are gods (Hephaestus) or legendary geniuses (Daedalus), then, as the centuries pass and stories change, wizards, and then scholastics (St. Albert the Great, Roger Bacon, etc.). The authority of those earliest, mythical creators comes from their own divinity; later, as creators become less legendary and more historical, it is their education or status in society …