This discussion was spawned in the Artificial Life commentary and I think it deserves thread of its own.
First of all Laws of Nature are those things which are observed over and over and over again without exception. We need not have physical theories to explain them. One such Law of Nature is the law of gravity. We have observed its effects countless times without exception. Mass is attracted to other masses. We don’t have a physical theory to explain the mechanism by which gravity works but due to empirical observations it is considered a law nonetheless. An exception may exist that disproves the law but until an exception is observed the law remains intact.
Another law that doesn’t get as much press is called the Law of Biogenesis – life comes only from life. It is supported by countless observations without exception. An exception may exist but until it is observed the law remains intact.
In Artificial Life pursuits there are two distinct approaches. One is a hands-off approach where natural environments are simulated, inanimate chemical suspects are put in contact, and we watch to see if anything interesting happens – if more and more complex self-replicators evolve without help. That has not yielded any fruit after many decades of trying. As the prior article pointed out even simulating this process in a computer has not borne fruit. In the words of principle investigators “something is missing”.
In the other approach we attempt to create artificial life with no holds barred. We throw all the intelligent intervention at the problem we can possibly come up with. This approach holds significant promise. And, here’s the kicker, this approach if and when successful is no exception to the Law of Biogenesis – life in that case still comes from life. It doesn’t come from inanimate chemicals dancing to the tune of law and chance alone. It comes together through the efforts of another living thing.
After getting into a discussion about Artificial Intelligence it occured to me there’s another law that is at work. Intelligence only comes from intelligence. This law, like all laws, is supported by countless observations without exception. Until an exception is observed the law remains intact.
Like the pursuit of Artificial Life the pursuit of Artificial Intellegence has two main approaches. One is a hands off approach where we throw together all the components we think are required for intelligence and see if anything interesting emerges by law and chance. This approach has not yielded any fruit. Nothing we could possibly call an independent intelligent agent has emerged.
On the other hand when we employ all the intelligent intervention we can come up with we have things with promise. Some even meeting a rudimentary definition of intelligent agency. The best example of this I’ve seen is the autonomous vehicles that completed the 2005 DARPA Grand Challenge. The challenge was to get a vehicle to self-navigate through 100 some miles of difficult desert terrain filled with obstacles. The vehicles were not allowed to contain a detailed course map. They were basically given the goal of getting from point A to point B with no more than the map a human driver might have. The vehicles were given sensory apparatus to gather input from the real world, they were given controls to alter their course and speed, and they were given the ability to model reality so that different control inputs could be tested in the abstract to see which result in furtherence of their goal. In other words as these vehicles go along they observe, predict, decide, and act accordingly. This is rudimentary intelligent agency. Granted it’s light years away from the sophistication of human intelligence but there’s nothing in principle that says we can’t keep improving the intelligence in these devices until they are that sophisticated.
But even if that happens, we do create an artificial intelligence similar in capability to our own intelligence, the law remains unbroken. It will be a case of one intelligence creating another intelligence. That will do nothing more than add further evidentiary weight to the proposed law that intelligence only comes from intelligence.