Opinion

Let's Do AI Right

David L. Heiserman

AI is extremely fertile ground for science & technology reporting. It takes a bit of study, but not a whole lot of digging to find the limitations of most science writers and bloggers.  I will be writing more on this particular topic, but I really want to make one thing clear:  Do not use the human brain as a model for AI theory and design. There's a good reason:  We haven't the foggiest idea how the human brain works. That being the case, you don't need a PhD in epistemology or critical thinking to see the absolute folly that is inherent in comparing computers with the human brain and the human brain with computers. We know how computers work, but we don't know how the brain works. The comparison is logically invalid and thus leads us away from fresh insight and a higher order of technological achievement.

In the 1970s, for example, I wrote a couple of books that proposed a model for autonomous intelligence--machines that programmed themselves according to the kind of external environment they encountered. Comparison to the human brain was totally avoided. Instead, my projects demonstrated a form of machine intelligence that had no dependence upon our own fragmented understanding of human intelligence. Do you understand the significance of that simple idea? Using today's technology for modeling the behavior of a totally autonomous intelligence gives us insight into a totally alien intelligence; and there is a good chance that knowledge can applied to a higher order of understanding of our own brains.

Here are three of my books on this topic and from that era. Although the technology is obsolete, the principles offered in these books remain valid today. You can get your own copies through Amazon for a dollar or less.

I have more recently published a Kindle book that summarizes the content of all the my machine intelligence and do-it-yourself robot projects:

Free phone apps

For your Alexa