There is little doubt that we are experiencing a resurgence of interest in artificial intelligence (AI). Most new household appliances are advertised as being “smart,” and we all carry in our pockets phones that can provide intelligent assistance and automation. Nearly every high-tech company is investing millions each year into AI research and development, with high-profile case studies like IBM’s Watson, Honda’s Asimo, and Google’s driverless car catching the imagination of the public. It is remarkable to think that, given the impressive nature of many of these modern breakthroughs, the field of AI is still relatively young. Markoff’s book seeks to trace many of the key strands of the history of AI from the first half of the 20th century to the modern day whilst simultaneously highlighting many of the ethical problems and unanswered questions.

Throughout the book, Markoff comes back again and again to the tension between those researchers who think that research within the field should focus on AI and those who think it should focus on intelligence automation (IA). This tension can be traced back to the viewpoints of two of the founders of the field, John McCarthy and Douglas Engelbart. Much like two great philosophers with different schools of thought, Markoff is keen to place the key players within the history of AI within either the McCarthy or Engelbart schools. The difference between these two schools of thought is whether or not the intelligent machines that we create should replace (AI) or augment (IA) human activity. This is why it is so useful for Markoff to dichotomize the field in such a way that provides a framework within which many of the ethical problems that occur as a result of AI/IA research can be discussed.

Each of the book’s nine chapters focuses on a single idea or key person within the history of AI. This personal approach to the history of the field makes the book eminently readable; in fact, once I started reading I found it very hard to put down. I did find Markoff’s style somewhat difficult to get used to, however. The train of thought within each chapter is often nonlinear; that is, Markoff will jump around in time and often discuss tangential topics that can make it hard to understand what the central argument of a section or chapter is. This point aside, Markoff’s skill as a journalist is evident throughout the book, with each chapter being meticulously researched. In terms of interest, the final chapters waned a bit when compared to the earlier chapters. This is epitomized by the hyper focus in the penultimate chapter on Apple’s Siri, an impressive technical advance that I am unsure warrants the amount of space and attention Markoff gives it.

One other concern that I have with this book is the lack of space that is given to recent advances within the field of deep learning (deploying neural networks on a large scale). Deep learning is mentioned in chapter 4, but the impact that this topic has had on AI and machine learning research would not be apparent to the uninformed reader. I have little doubt that the advances we are seeing in the field as a result of deep learning represent a paradigm shift in the way that researchers and practitioners approach AI and IA. It will be interesting to see if any future editions of this book will address these advances in more detail.

Overall, this book weaves together multiple threads of thought and argument relating to the history of AI and IA research. It provides a panoramic history of the field along with a discussion of the conflict between AI and IA. Alongside these two threads is woven the question of the ethical implication of automation and its social and economic impact. I would recommend this book to anyone interested in understanding the history of the field of AI and also those who want to understand the ethical implications of developing “machines of loving grace.”

More reviews about this item: Amazon, Goodreads