Artificial Intelligence: A Guide for Thinking humans – Melanie Mitchell
The author is a professor of computer science specializing in artificial intelligence (AI). She is also the author of several other books on the subject. Regardless of her credentials this book is unique in its approach to the subject. It is written for the layperson who is not a computer scientist. Even I could read it and understand the concepts she discusses.
Ms. Mitchell provides a revealing history of artificial intelligence from its earliest days in the mid-1950s through the digital computer revolution of the ‘70s and ‘80s and into today’s efforts to build systems as intelligent as humans. She discusses symbolic and sub-symbolic systems, neural networks, as well as various other AI approaches, and the shortcomings and successes of each. Don’t understand what these systems are or even the terms? You don’t need to. The author does an outstanding job of describing and explaining every concept in terms clear enough for average non-technical people to understand.
The author dedicates substantial effort to answering the most urgent questions about AI in today’s reality: How really intelligent are the best AI programs? How do they work? What can they actually do? When do they fail? How humanlike do we expect them to become? And how soon are they likely to surpass human intelligence?
Ms. Mitchell explores the profound disconnect between the media headlines that describe miraculous breakthroughs in machine learning and capabilities of AI programs, and the actual performance of these systems. She praises the achievements of AI engineers and scientists, especially in the areas of machine vision as well as voice recognition and language translation programs. These seem similar to human abilities. She affirms that these machines can do remarkable things, and can be “trained” to be much better at some specific tasks than humans. However, no systems are yet able to transfer intelligence learned in one task to accomplishing a different task. That requires “general intelligence” which so far only humans have.
The last chapter was especially interesting to me. Ms Mitchell gives her personal perspective by answering a short series of questions about the future of AI; future of autonomous vehicles is an example. Throughout the book the author gives the reader a clear sense of what the field has accomplished and how much further it has to go. Great read!!