The Brain Race

Back in the 1930's Leo Szilard foresaw the possiblity of a super bomb based on nuclear fission. This is talked about in the fascinating book The Making of the Atomic Bomb by Richard Rhodes. Szilard became an advocate for research into the underlying physics. He was extremely afraid of an enemy nation creating such a bomb before America. The reasons are obvious. If an enemy nation were to create enough super bombs, and detonate them against the U.S. and its allies, the U.S. might have no choice but to surrender.

Szilard was instrumental in alerting the U.S. government to the potential and threat of such an atomic bomb. Eventually the U.S. government went into action and started the Manhattan Project which culminated in the creation of the the atomic bomb, before any other nation on earth could do so. This was a race for supremacy, and the U.S. had won hands down.

In the 1960's the Space Race came about. This was a race for technological superiority in the ability to loft payloads into outer space. The stated goal was putting a man on the moon before 1970, however it is clear in retrospect the main concern had to have been national defense. If an enemy power could manage to put weapons in space first, it could conceivably attack any place at any time, and could at the same time prevent other countries from being able to put anything in space.

Once again the U.S. had won the race. This one was close though, the Soviet Union had been trying for the prize as well, but some unfortunate setbacks (like a massive explosion of their equivalent of the Saturn-V rocket) left them out of the running. At the time the Soviet Union had downplayed their effort to win the space race. But looking back it is clear they were doing their best to win it.

In both cases the U.S. became involved in a very expensive technological race. In both cases the U.S. could have easily been convinced this was a fight to the death -- if the enemy wins the race, they'd waste no time in taking over the U.S. and imposing their way of life.

In both cases, having won the race in question, the U.S. simply went back to life as usual. The U.S. in its own right didn't then conquer the world -- even though such capability was well within its power. The U.S. woke up, went to work, and when the job was done went back to sleep again.

Right now, the U.S. faces another threat just like these previous two. It is a threat that has the same crazy science-fiction feel about it. It's fantasy. However a nuclear super bomb was just science fiction fantasy before it was created. Putting a man on the moon was the stuff of science fiction as well. But once both projects became national priorities, success was achieved in the blink of an eye in a historical sense. Well under 10 years.

The new threat is the creation of the first artificial brain. The effort to create it could be called the Brain Race.

Specifically an artificial brain is a mechanism that is intelligent, self aware, has free will, has consciousness, and is at least as intelligent as an average human. You can talk to it, and it answers like a person would. It can be taught to drive a car, write books, write computer software, aim a gun, and pull the trigger. What form it might take isn't very important. The issue is creating the first one. Once that is created, technology can be applied to refinements and improvements. What is lacking is not resources or computing power, but an understanding of the underlying principles in how the human brain functions.

Once a functioning artificial brain is made, the technology can be improved to make mass produced copies of it viable. Perhaps the first one needs the resources of a supercomputer (or several) to operate. However the second one will be vastly more efficient, and will be based on custom machinery that has been optimized for the specific principles that give rise to intelligence. Given the basic understanding, reducing the resource requirements to bare minimums will be trivial. One can imagine in a very short time producing a device the size of a cell phone that is just as intelligent as a human.

Where is the threat? The threat is that such mass produced intelligent machines can be the basis of an unlimited number of willing, intelligent slaves. Placed in control of robotic machinery, they could then be used to operate factories to produce yet more robotic machinery, in an exponential growth scenario, limited only by the laws of physics. Refinements in the underlying principles will allow for the creation of artificial brains very much more intelligent than we are. These can in turn guide the development of still more intelligent brains. The construction of more adept and agile robots would be perfectly viable.

Imagine a robot that looks like a human, yet has reflexes and strength far beyond human capacity. Imagine enough of them operating as willing slaves to a hostile nation. Imagine such robots having built into them enough high explosive to level a shopping mall. Imagine them coming in droves on the beaches, making their way to populated areas, then exploding when they get near groups of people.

They wouldn't even need to self destruct. They could just kill Americans in whatever convenient manner they want -- clubbing, strangulation, bullets. Whatever, they'll be adaptable. It won't matter if the army is mobilized against them. Kill one and there are more on the way. They're all clones of one original, mass produced like so many toasters.

This is just one scenario. It's not even a clever one. Without much imagination lots more scarier scenarios can be envisioned. The point is that with the ability to create intelligent machines, a nation, group or even an individual would be able to in short order do anything they would like with the planet and all the people living here, and nothing could be done to resist.

Somewhere soon the first truly intelligent artificial brain is going to be created. Or perhaps discovered. I just hope whoever makes that first discovery is benign, as the U.S. has proven to be on two previous occasions.

David Ashley
March 2007

Access count (e^i)