MAIN IDEA:

This book is about intelligence. It explores the idea of intelligence in humans and machines. The main point is that humans must find a way to control AI, even when it acquires superior functionality in processing existing information and generating new one, such as ideas, technology, and art.

MY TAKE ON IT:

This book is a pretty good presentation of the concept of intelligence as a more or less abstract idea. It also includes detailed descriptions of current debates and provides some suggestions for dealing with AI. I think the whole problem is a bit overblown because of the dominant concept of humans as purely biological objects as we are now and have always been. However, if one looks at a human as a self-directing information-processing object deeply embedded into the network of interactions with other humans, the problem kind of disappears. Humans are all different from each other in their DNA, cultural background, and behavior. So, why is it a problem if humans create a silicon-based self-directing information-processing object with all human characteristics except for the biological base? I do not think it will happen because nobody needs it, maybe except for some controlled experimentation. Beyond that, AI will be used as all other machines: to process materials and information to satisfy biological humans’ material and intellectual needs. The key here is the notion of SELF, which no AI computer could obtain without being “raised” in the same way as human beings are. Surely, it could be done much faster, but the nature of the process has to be the same.