We might even imagine that as a robot becomes more intelligent and self-aware, its brain might become sensitive enough to undergo harm if it were forced to do something needlessly embarrassing or undignified. Consequently, the Second Law of Humanics would be:
2-A human being must give orders to a robot that preserve robotic existence, unless such orders cause harm or discomfort to human beings.
The Third Law of Robotics is designed to protect the robot, but from the robotic view it can be seen that it does not go far enough. The robot must sacrifice its existence if the First or Second Law makes that necessary. Where the First Law is concerned, there can be no argument. A robot must give up its existence if that is the only way it can avoid doing harm to a human being or can prevent harm from coming to a human being. If we admit the innate superiority of any human being to any robot (which is something I am a little reluctant to admit, actually), then this is inevitable.
On the other hand, must a robot give up its existence merely in obedience to an order that might be trivial, or even malicious? In “The Bicentennial Man,” I have some hoodlums deliberately order a robot to take itself apart for the fun of watching that happen. The Third Law of Humanics must therefore be:
3-A human being must not harm a robot, or, through inaction, allow a robot to come to harm, unless such harm is needed to keep a human being from harm or to allow a vital order to be carried out.
Of course, we cannot enforce these laws as we can the Robotic Laws. We cannot design human brains as we design robot brains. It is, however, a beginning, and I honestly think that if we are to have power over intelligent robots, we must feel a corresponding responsibility for them, as the human character in my story “The Bicentennial Man” said.
Cybernetic Organism
A robot is a robot and an organism is an organism.
An organism, as we all know, is built up of cells. From the molecular standpoint, its key molecules are nucleic acids and proteins. These float in a watery medium, and the whole has a bony support system. If is useless to go on with the description, since we are all familiar with organisms and since we are examples of them ourselves.
A robot, on the other hand, is (as usually pictured in science fiction) an object, more or less resembling a human being, constructed out of strong, rust-resistant metal. Science fiction writers are generally chary of describing the robotic details too closely since they are not usually essential to the story and the writers are generally at a loss how to do so.
The impression one gets from the stories, however, is that a robot is wired, so that it has wires through which electricity flows rather than tubes through which blood flows. The ultimate source of power is either unnamed, or is assumed to partake of the nature of nuclear power.
What of the robotic brain?
When I wrote my first few robot stories in 1939 and 1940, I imagined a “positronic brain” of a spongy type of platinum-iridium alloy. It was platinum-iridium because that is a particularly inert metal and is least likely to undergo chemical changes. It was spongy so that it would offer an enormous surface on which electrical patterns could be formed and un-formed. It was “positronic” because four years before my first robot story, the positron had been discovered as a reverse kind of electron, so that “positronic” in place of “electronic” had a delightful science-fiction sound.
Nowadays, of course, my positronic platinum-iridium brain is hopelessly archaic. Even ten years after its invention it became outmoded. By the end of the 1940s, we came to realize that a robot’s brain must be a kind of computer. Indeed, if a robot were to be as complex as the robots in my most recent novels, the robot brain-computer must be every bit as complex as the human brain. It must be made of tiny microchips no larger than, and as complex as, brain cells.
But now let us try to imagine something that is neither organism nor robot, but a combination of the two. Perhaps we can think of it as an organism-robot or “orbot.” That would clearly be a poor name, for it is only “robot” with the first two letters transposed. To say “orgabot,” instead, is to be stuck with a rather ugly word.
We might call it a robot-organism, or a “robotanism,” which, again, is ugly or “roborg.” To my ears, “roborg” doesn’t sound bad, but we can’t have that. Something else has arisen.
The science of computers was given the name “cybernetics” by Norbert Weiner a generation ago, so that if we consider something that is part robot and part organism and remember that a robot is cybernetic in nature, we might think of the mixture as a “cybernetic organism,” or a “cyborg.” In fact, that is the name that has stuck and is used.
To see what a cyborg might be, let’s try starting with a human organism and moving toward a robot; and when we are quite done with that, let’s start with a robot and move toward a human being.
To move from a human organism toward a robot, we must begin replacing portions of the human organism with robotic parts. We already do that in some ways. For instance, a good percentage of the original material of my teeth is now metallic, and metal is, of course, the robotic substance par excellence.
The replacements don’t have to be metallic, of course. Some parts of my teeth are now ceramic in nature, and can’t be told at a glance from the natural dentine. Still, even though dentine is ceramic in appearance and even, to an extent, in chemical structure, it was originally laid down by living material and bears the marks of its origin. The ceramic that has replaced the dentine shows no trace of life, now or ever.
We can go further. My breastbone, which had to be split longitudinally in an operation a few years back is now held together by metallic staples, which have remained in place ever since. My sister-in-law has an artificial hip-joint replacement. There are people who have artificial arms or legs and such non-living limbs are being designed, as time passes on, to be ever more complex and useful. There are people who have lived for days and even months with artificial hearts, and many more people who live for years with pacemakers.
We can imagine, little by little, this part and that part of the human being replaced by inorganic materials and engineering devices. Is there any part which we would find difficult to replace, even in imagination?
I don’t think anyone would hesitate there. Replace every part of the human being but one-the limbs, the heart, the liver, the skeleton, and so on-and the product would remain human. It would be a human being with artificial parts, but it would be a human being.
But what about the brain?
Surely, if there is one thing that makes us human it is the brain. If there is one thing that makes us a human individual, it is the intensely complex makeup, the emotions, the learning, the memory content of our particular brain. You can’t simply replace a brain with a thinking device off some factory shelf. You have to put in something that incorporates all that a natural brain has learned, that possesses all its memory, and that mimics its exact pattern of working.
An artificial limb might not work exactly like a natural one, but might still serve the purpose. The same might be true of an artificial lung, kidney, or liver. An artificial brain, however, must be the precise replica of the brain it replaces, or the human being in question is no longer the same human being.
It is the brain, then, that is the sticking point in going from human organism to robot.
And the reverse?
In “The Bicentennial Man,” I described the passage of my robot-hero, Andrew Martin, from robot to man. Little by little, he had himself changed, till his every visible part was human in appearance. He displayed an intelligence that was increasingly equivalent (or even superior) to that of a man. He was an artist, a historian, a scientist, an administrator. He forced the passage of laws guaranteeing robotic rights, and achieved respect and admiration in the fullest degree.