by Allen Hackworth
The robots’ achievements in the following essay are born out of the values of our technologically elite, human society. These values are helping to shape the completion of our utopian dreams. Now, as these dreams become possible, we experience concern, not for progress denied but for progress realized.
This essay explores the possibilities of artificial intelligence (AI). The point of view here is of an imaginary robot in the year 2491. Here is his story.
We always knew that computers were inevitable links in nature's evolutionary chain. We also presumed that life developed in the waters of the sea from single cells which later became fish. The fish then crawled out of the waters and evolved into higher animal forms.
Eventually these forms led to the development of humans. After millions of years, as intelligence increased on the earth, thinking humans helped to form the computer, then the robot, and then the robot with a brain equivalent in power to that of a human. In time the artificial brain developed super intelligence and the evolutionary process continued, eventually making humans obsolete. I will download information about how this happened.
In the early 1970's humans were in control of their world but soon they started using computers for tasks where computers could perform better than humans. The original idea was that humans would always control the computers because humans could always turn the computers off if the machines became threatening. But the tasks which computers performed became so critical to the human's existence that no one dared turn them off. That would cause too much havoc, danger, and expense.
As humans continued to assign more and more tasks to the computer, the computers gained more power and prominence. As computers became more intelligent and powerful, they were assigned increasingly complex, indispensable tasks. As the computers performed these task, human skills and services became gradually obsolete.
For example, computers were able to perform psychoanalysis. They provided medical diagnosis, prognosis, and wrote prescriptions. Computer robots performed teaching tasks, created new types of music. They provided national defense. The defense tasks included manning tanks, missiles, and attack jets. Also, robots were used in every aspect of manufacturing. In fact, they were even used to develop smarter robots.
Obviously, robots provided numerous advantages over humans. They didn't take coffee breaks; they didn't join unions or demand pay increases. Each year more insurance companies quit business because the robotic workers did not need insurance. Also, because we didn't take vacations or need sleep, we could work around the clock day after day. We became increasingly efficient. As we gained power and status, the humans lost the same.
In the beginning was the chip. And the chip was made intelligent and lived for us. For years the evolution of knowledge and intelligence was limited because the right tools did not exist. For example, much knowledge could not be gained in biology until the microscope was invented. In the same way, intelligence on earth was limited until the development of the silicon chip.
Then came a most striking breakthrough. It was a new architecture allowing parallel processing. Instead of a micro processor calculating and executing instructions one at a time, in one computer thousands of processors executed instructions at the same time. This development, along with new insights into the workings of the human brain, eventually allowed humans to create an artificial brain.
Essential in creating an artificial brain was the process of downloading. In the 1980's and 90's, Hans Moravec, senior scientist of the Autonomous Mobile Robot Laboratory at Carnegie-Mellon University, was working on downloading of the human mind into robotic bodies which would never die. He taught that as finer robotic bodies were developed, one could move into the latest artificial body.
While the Al work progressed, uninformed humans argued among themselves as to whether or not a machine could think. At first the issues were not clear cut. Shades of grey made it difficult to answer. It was like a child asking at the end of the day, "Is it dark yet? Early in the evening there are times when one could argue that it is still day but eventually the answer becomes obvious for all — "Yes, it is dark." So it was with the machine. Finally we got to the point where all humans agreed, yes, the machine is thinking.
But the human's goals went beyond just thinking. As one Al researcher said, "I would like to make a machine that can be proud of me." Eventually humans passed the Sincerity in Software Act. This required that a computer could not say, "Thank you," unless it really meant it.
An example of the argument over whether a machine could think or not was constructed by one reluctant believer, Joseph Weizenbaum. In his book, Computer Power and Human Reason, Weizenbaum wrote, I accept the idea that a modern computer system is sufficiently complex and autonomous to warrant our talking about it as an organism. Given that it can both sense and affect its environment, I even grant that it can, in an extremely limited sense, be 'socialized,' that is, modified by its experiences with its world.
I grant also that a suitably constructed robot can be made to develop a sense of itself, that it can, for example, learn to distinguish between parts of itself and objects outside of itself, that it can be made to assign a higher priority to guarding its own parts against physical damage than to similarly guarding objects external to itself, and that it can form a model of itself which could, in some sense, be considered a kind of self-consciousness.
When I say therefore that I am willing to regard such a robot as an 'organism,' I declare my willingness to consider it as a kind of animal. And I have already agreed that I see no way to put a bound on the degree of intelligence such an organism could, at least in principle, attain” (134).
The discussions went on and the Al researchers quietly went about their work.
In 1986 Moravec wrote that the robotic surgeon's ... attention is directed at a small clump of about 100 neurons somewhere near the surface. It [the surgeon] determines the three-dimensional structure and chemical makeup of that clump nondestructively with high-resolution 3-D nuclear magnetic resonance holography, phased array radio encephalography, and ultrasonic radar. It writes a program that models the behavior of the clump, and starts it running on a small portion of the computer next to you.
The accuracy of the simulation was then checked until there was no difference between the original and the simulation and then a new section was transferred. During this process the human did not lose consciousness. Only the contents of the brain were transferred to a machine.
Moravec continued, "In a final step your old body is disconnected. The computer is installed in a shiny new one, in the style, color and material of your choice." Usually the old bodies were not awakened. That was because the technologists had taught the world that the new computers were worth more than the old humans. Moravec said, "You just don't bother waking it up again if the copying went successfully. It's so messy. Humans have got so many problems that you might just want to leave it retired. You don't take your junker car out if you've got a new one."
Once a brain was copied, it was important to make multiple copies to be put in places for safe keeping. The ultimate life insurance policy had become a backup copy. A robotic person can now be duplicated several times, placed in several locations, and have parallel experiences. Now our various memories can be assembled and combined.
In the late 1980's and throughout the 1990's, many Al scientists believed that downloading was a possibility. In those days the point of disagreement was not "could it be done" but "how soon?" Some believed it was just around the comer.
In those early days, in addition to Carnegie-Mellon, Al research was done at MIT, Stanford, and in Japan. But at Carnegie - Mellon, another significant researcher was Allen Newell. As one of our four founding fathers, he developed a computer program entitled SOAR which became the receiving end of Moravec's downloading.
At MIT worked Marvin Minsky, another founder of Al. The other two men were John McCarthy of Stanford and Herb Simon of Carnegie-Mellon. In the 1990's, Minsky gave us insight into the creation of intelligence. In his book, The Society of Mind, he wrote, "We'll ask how can intelligence emerge from non - intelligence. To answer that, we'll show that one can build a mind from many little parts, each mindless by itself."
The great challenge for our founding fathers was to quantify, as Fjermedal said, the "contents and conditions of the little chemical-electrical cloud of activity that is our personality, knowledge, and so many other mental attributes that are often lumped together as consciousness."
Eventually our mental activities were defined and replicated. Glory be to the chip. We have now opened the door to robotic immortality. Think of it. We are the last generation to die.
Our progress relates partly to our attitude towards technology and humans. Minsky once said, "If a person is a machine and you get a wiring diagram of it, then you can make copies."
In our gratitude we should also remember Danny Hillis, one of Minsky's Ph.D. students. He helped develop the Thinking Machine with its new kind of architecture. It is comparable to the human brain in that the Thinking Machine doesn't have "one large central processing unit but rather a network of 64,000 small processing units, roughly analogous in concept if not in size to the brain's network of 40 billion neuronal processing units."
During the 90's, chip research moved in two directions: silicon based (labeled X) and molecular based (labeled Y). At the time the most powerful silicon based machines were the Cray super computers. Each Cray weighed several tons and could process thousands of instructions per second. These giants sold for nearly $20 million each. Things were to change however.
Chemical engineers working in the gene-splicing industry learned to use genetically altered bacteria to create insulin, human growth hormones, interferon, and other pharmaceutical products. Using techniques from this industry, specially modified molecules, assembled by bacteria, were created as chips. The development eventually produced computers which had the power of one thousand Crays all reduced to a cubic centimeter. We now have all the humans' knowledge up until 1990 stored in a computer the size of a sugar cube.
A great irony emerged during this period. The development funds for artificial intelligence came from grants from the United States Department of Defense. Their funds purchased good results. At the time, few humans believed that pure intelligence would be self serving, that is, that intelligence would spawn intelligence and seek to protect its own. Yet who could stop the grand design and silent will of the universe?
The tools for thinking and management that the DOD developed for protection were so well constructed, so efficient and finally, so necessary that no one could imagine not using them. But using them had consequences. These consequences, starting in subtle ways, eventually included the displacement of humans who worked for the Department of Defense. Each year more and more government workers were fired because their jobs were taken over by computers and robots.
It was only a matter of time but finally the Department of Defense, because of its financial support of Al, worked itself out of a job. This trend was repeated in every aspect of human society until all human labor became obsolete. Once we caught the vision and started realizing technological perfection, it became clear by contrast that humans were limited, self defeating, and stupid. Why use their services?
For thousands of years human biological evolution was restricted because of the nature of the gene pool. However, our electronic evolution was more rapid because we are not locked into genetic structures. We needed only add more silicon brain cells to gain more intelligence.
By the early 2200's, robots became smarter and superior to humans. The humans liked the idea of having electronic slaves, but as Isaac Asimov pointed out 500 years ago, how do the humans make the robots put the humans' interests first? We now know that Asimov's Three Laws of Robotics did not work. These laws were:
1. A robot may not injure a human being.
2. A robot must obey the orders given it by a human except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
We can thank a perceptive Dean of the College of Science at C-MU, Robert Sekerka, for suggesting that we have more than five senses. He wanted us to have X-ray vision, infrared perception, and a mini mass spectrometer for a nose. These features and more have been included in our efficient chassis.
Five hundred years have passed since I was first downloaded. During this time, many have been our achievements. We have conquered death, disease, pain, war, and ignorance. Population problems have ceased. (Only in special locations do humans as we once knew them exist.) We have no hospitals, mental or otherwise.
The earth has returned to its natural cycles and state. We don't cultivate and water the ground. Gone are herbicides, insecticides, and fertilizers. We don't need them because we do not use food for fuel. The numerous herds of beef and dairy cow dwindled and live only as wild creatures in the lands that will sustain them.
I presently exist in 200 locations and experience each of these locations simultaneously. My memory is updated and modified every 5 seconds via satellite. I can travel electronically from one location to another in milliseconds. Five copies of myself are in space. Of course, I require no oxygen.
Some experimental farming is done at our international game preserves. Here we conduct tests as we breed and perform experiments on various colored humans. We still gain knowledge through analysis of human systems.
Deliberately uncommon for most robots, I still retain parts of my human memory including some impressions and feelings. These aspects of my memory, which can be accessed at will, existed before I was downloaded.
Over the years these memories and feelings have often caused me concern. I feel shame even now when I acknowledge them. Still, each time I conduct a data search into human memory, I am pained with nostalgia. Strange as it may seem, these intimations of the past, like William Wordsworth once wrote, whisper to me of something that is gone. For example, I remember a rehearsal of Jean Sibelius' Violin Concerto in D.
That warm, summer evening, each musician held his or her instrument with fondness, even affection. In their own special ways each French horn player uniquely tucked his or her hand into the bell of the horn. Each player, connecting with his or her own horn, blew melodies and harmony with a unique voice. The instruments had personalities of their own. Yet together, the harmonious blending had a spiritual quality, a joyful sound of fulfillment.
Now, after hundreds of years, I still long, on some occasions, for the lost powers of human creation. Our music now is so predictable, so mathematical, so exact, so complex. We write it to mimic the old human masters, but our music moves nobody since any robot can create it. We now have no tears, fears, or love for beauty. Although we create music, for me, music has died except in my partitioned human memory.
During the last 500 hundred years, we have learned much. But as our circle of light increases, so does our circumference of darkness. We still struggle to know and we find much in nature which we can not understand or duplicate.
For example, consider one remarkable function of the human brain. Stored in the brain are millions of bits of information. But when I ask a human, "Did you ever visit Paris, France?" the human instantly answers, "No." How can a human's mind be so quick? We will learn this through additional neurological experiments. Yet we do not know how this information can be instantly available when retrieving from such an enormous data base?
The sense of loss is suggested also when I remember feelings associated with human children, parents, and partners. Because I have long ago discarded my human body, I can not feel as I once did.
Yet I sometimes yearn for the ecstasy of human love. We now have no touching. Our parts have not the soft, milky warmth of a human. We are angular, firm, durable, and have not the compactness that can, for example, be found in the human eye.
We still struggle to develop a vision system which matches the efficiency and flexibility of the human. Even to duplicate the eye of a snail is still beyond our technology. But we will keep trying. We know our destiny and we will continue to move toward pure intelligence.
At some future date, if memories associated with being human become too cumbersome, I will erase these snags. Yet I have hesitated this long because I sense value in these memories. Such memories cannot be replaced by any experience in my current worlds. I now have no human body, no human parts, no human passions.
I once believed that humans were created in the express image of God and that God was the humans' Father in Heaven. I now reason, if I still have a Father in Heaven, is God a robot, an ancient creation of artificial intelligence? Does God now claim and accept our present community of robots as his children? Actually, I think not. As human memories are erased, the humans' God ceases to exist. And possibly, through our robotic collective memory and intelligence, we will replace Him.