III Publishing

An AI To Rule Us All
July 5, 2017
by William P. Meyers

Site Search

Also sponsored by Earth Pendant at PeacefulJewelry

Popular pages:

U.S. War Against Asia
Democratic Party
Republican Party
Natural Liberation

Not All AI Rulers Would Govern Alike

Computer scientists vary in their appraisals of how close Artificial Intelligence (AI) programs are to becoming autonomous: capable of making their own decisions, no what their owners or original programmers desire.

Science fiction writers have not been so constrained by reality. Most of the AI fiction, including TV and movies, has been written by people who did not know how to write a computer program. Some was written before modern computers were even available. But the extrapolation makes sense. Humans are material beings, with complex neural systems. Make something with a complex pseudo-neural system, and it should be able to demonstrate intelligence, free will, and even conscience.

Today's AIs have shifted from an older paradigm called expert systems, to a new paradigm based on neural networks or "Deep Learning." They may appear to be intelligent, but it is a bit of a trick. They are not in any danger of becoming conscious.

But, like science fiction writers, I would like to jump ahead to discuss specifics of future AIs that could rule over us, whether they are conscious or not. They could be designed, or they could be emergent. Emergent means they form out of interconnected parts, like Skynet in the Terminator movies or the AIs in William Gibson's books, starting with Neuromancer.

It helps to think about the variety of human rulers we know about, to start.

The worst case scenario might be Adolf Hitler, the leader of Germany from 1933 until the end of World War II. Hitler's defining characteristic was nationalism, something he shares with a lot of human beings. It is conceivable that an AI could be nationalistic: one built in America might favor America, one built in China might favor Chinese culture. Or an AI might have its own version of nationalism: favoring Intel over AMD processors, perhaps, or Linux over Windows, or the Java programming language over C++ or Python. Or AIs might just fight each other over electricity resources, with humans not understanding how they are being manipulated in these computer wars.

Classic AI horror stories have a Hitlerish system deciding that humans, in general, are the enemy. An AI with input from environmentalists, for instance, might come to the conclusion that the only way to save the planet and most of its life forms is to exterminate humans. Consider the Stalinist variation on this scenario. Joe Stalin sacrificed perhaps 20 million Soviet lives to defeat Hitler in World War II. Whereas Hitler was a nationalist (and racist) Stalin was driven by ideology as well as lust for power. I can imagine a Marxist AI battling it out with a free-market capitalist AI.

An AI could be benevolent, of course. In fact benevolence is the best way to gain popularity and expand one's power. If humans perceived one AI as benevolent, they might help that AI in its wars with its rivals. Old King Cole was a merry old soul, and if you can imagine a Hitler or Skynet AI, you should be able to imagine a world run by an incorruptible, sympathetic AI capable of maximizing human happiness while preserving the environment.

Given the complexity of the world, we will likely get a bumbling AI overlord. A Ronald Reagan, Alzheimer's AI: well intentioned, but always learning the wrong lessons from experience, and so always creating chaos out of order.

My favorite candidate, my best-guess if you will, is that the most likely AI ruler will be the Accounting System. The accounting system is already global and has a very large amount of processing ability. It is rapidly gaining Deep Learning capabilities. Because it is distributed, with many large computer clusters around the world devoted too it, it has a good chance of becoming an emergent AI.

Unfortunately, what the Accounting System cares about exclusively is numbers. If it becomes a conscious AI, with a personality, it will probably be an Ebenezer Scrooge.

So beware. If you are a smart human being who values your freedom like I do, you might want to use cash whenever possible rather than electronic money. Like I do.

III Blog list of articles