by Dave Moore, CISSP
11/12/2023
Continuing from last week, from the session I attended in Las Vegas last August when I was at the AI4 2023 international conference on artificial intelligence, which featured Mark Brady, PhD.
Last week, we looked at how Dr. Brady covered the “First Singularity.” This week, “the Second Singularity.”
The Second Singularity is the one that freaks people out the most, the one that’s in the doomsday science fiction movies. You’ve seen Arnold Schwarzenegger in “The Terminator,” right? We’re talking about that type of “singularity.” While it’s easy to dismiss such concerns as conspiracy-theory doom and gloom, it’s not so easy to dismiss the fact that it could happen. It is, in reality, very doomy and gloomy.
“The Second Singularity,” said Dr. Brady, “is the one you’re already familiar with and this is where AI reaches a point where it sort of takes off and becomes uncontrollable and irreversible, resulting in unforeseeable changes to human civilization.”
How can we predict when the Second Singularity will occur? “The second Singularity is going to occur when machine intelligence completely eclipses humans so it has to do everything that humans do, and then some.”
Brady then compares human versus traditional “machine intelligence.” Humans have a fallible memory; we forget things. Machines have infallible memory. They don’t forget things. Human memory is associative and machine memory is concrete. Humans have fallible calculation skills and fallible logical processing, machine calculation and logical processing is infallible.
Two areas stand out where, in our current world, humans still have a superior advantage: analogical reasoning and associative memory. So far, AI is not capable of either one.
“Analogical reasoning is the ability to make analogies,” Dr. Brady said. “Associative memory is memory where when you learn something new, you hang it on a framework of things that you already know. This is very important because it’s the basis of analogical reasoning, so humans are able to carry out analogical reasoning and machines traditionally are non-analogical.”
An example of analogical reasoning would be the development of the theory of electromagnetism. Maxwell and other physicists of the time were inspired by the work of Newton and his theories, but there’s no logical connection that should lead someone to put together the theory of electromagnetism. “You can’t take Newton’s theories and then by deduction produce the theories of electromagnetism; that is done by analogical reasoning.”
Intuition is another thing that’s very important, and goes hand-in-hand with analogical reasoning, but where do analogies start?
“What is the first thing that we we apply an analogy to?” Dr. Brady asks. “It can only be one thing and that is physical experience. We need to have physical experience in order for a machine to be intelligent to the level of the Second Singularity.”
“Intuition is another thing which is dependent on physical experience. A study of Nobel Prize winners showed that 70 percent felt their discoveries were based on intuition. You know how many axioms in mathematics are based on intuition? All of them, so, intuition is the real thing.”
All that’s missing is for robots to have physical experiences like humans do, which enable humans to have intuition and analogical thinking. For that, robots need something akin to our human five senses. See. Hear. Taste. Touch. Smell.
“Here’s a prediction: the Second Singularity is going to occur when machine intelligence is capable of analogical reasoning and has a robotic embodiment which will allow it to get that “real world” experience,” Dr. Brady says.
I can’t wait.
Dave Moore, CISSP, has been fixing computers in Oklahoma since 1984. Founder of the non-profit Internet Safety Group Ltd., he also teaches Internet safety community training workshops. He can be reached at 405-919-9901 or www.internetsafetygroup.org