Book Review: The Limits Of Artificial Intelligence

The Limits Of Artificial Intelligence, by Jacob T. Schwartz

As someone who often ponders questions of design and intelligence from a variety of different angles [1], I am often intrigued by the problem of computers and artificial intelligence.  After all, we gain a greater understanding of the impressiveness of our design when we engage in the process of designing ourselves.  Knowing the difficulties of dealing with trade-offs, for example, helps us to appreciate the subtlety of our own design and encourages us to stop considering every perceived shortcoming of that design as evidence of imaginary naturalistic processes.  Not surprisingly, this short book has a lot to say about such trade-offs, pointing out that there are real limits to computational efficiency when any constraints are relaxed in programs.  Also, the attempt to program genuine linguistic understanding has helped convince many even at the time of this short book being written in 1980s of the immense ambiguity and difficulty of human language.  Even thirty years after being written this particular book has much to say about the limitations that scientists and programmers still struggle with in the nascent field of artificial intelligence, even though the achievements made are still remarkable.

The contents of this short book are basically one moderately long essay of material that was originally written for the Encyclopedia of Artificial Intelligence but was also deemed of sufficient interest to be reprinted independently as a book.  To be sure, much about the book is dated and were it written today there would undoubtedly be discussion of the way that artificial intelligence has moved from chess to Jeopardy to the point where computers are now assisting in the writing of songs, which speaks to the decline of contemporary songwriting just as much as it does to the advance of computing.  The article is made for someone with a great degree of interest in applied mathematics, and who can at least follow along with the proofs included in the book.  Little concession is made to the reader by using anything less than the most advanced and technical vocabulary, but despite this there are a few matters of great interest that are discussed that many readers will find intriguing.  Among these matters is a naive belief that there is no difference between the intelligence that we can program and the intelligence that we possess, a belief that leads him to some obvious concerns about how artificial intelligence will be used and what that means for us.  Likewise, aside from some very advanced levels of imperfection, the author does not see any real fundamental limitations to the development of artificial intelligence and to a much deeper understanding within the field as to what it is about.

The author, moreover, seems to be filled with a sense of fear about two matters that many readers will find themselves similarly concerned about with regards to increases in robotics and artificial intelligence.  The first is that if robots and computers can be trained to do everything better than we can, there will be no dignity or employment or utility to human life, all of which will make it somewhat pointless to live except for the point of enjoying ourselves for a brief time on this planet.  The second is that since the author sees no difference between natural and artificial intelligence, the promise of exploitation of artificial intelligence for immoral ends means that there will be no protection against the exploitation of natural intelligence for the same ends.  While it is heartwarming and touching that the author expresses proper skepticism of the powers that be in our present evil age, and also that he considers the less robust intelligence of animals worth protecting on account of their possession of at least some capacity for intellect, the author’s lack of grounding in a proper theology of creation has denied him territory by which to deny the legitimacy of those who come from a philosophy of might makes right.  As is often the case in our world, our technological advancement has not in any way been matched by moral advancement, and even the author can see tragic implications of that.

[1] See, for example:

https://edgeinducedcohesion.blog/2012/09/12/book-review-i-robot/

https://edgeinducedcohesion.blog/2016/01/15/the-mind-is-not-a-machine/

https://edgeinducedcohesion.blog/2014/07/07/signs-of-mind/

https://edgeinducedcohesion.blog/2016/11/11/book-review-think-like-a-programmer/

https://edgeinducedcohesion.blog/2016/08/20/book-review-your-code-as-a-crime-scene/

About nathanalbright

I'm a person with diverse interests who loves to read. If you want to know something about me, just ask.
This entry was posted in Book Reviews and tagged , , , , , . Bookmark the permalink.

4 Responses to Book Review: The Limits Of Artificial Intelligence

  1. Pingback: Book Review: The Great International Math On Keys Book | Edge Induced Cohesion

  2. Biblio Phile says:

    Artificial intelligence is becoming more prevalent in everyday life whether we like it or not, so it’s best to learn all we can about its possibilities and dangers. I’ve put together a few selections that help folks get a handle on AI:
    http://booksto.co/learn-about-artificial-intelligence/

  3. Pingback: Book Review: The Little Book Of Mathematical Principles, Theories, & Things | Edge Induced Cohesion

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s