by Michio Kaku
One robot nurse that already exists is the RP-6 mobile robot, which is being deployed in hospitals such as the UCLA Medical Center. It is basically a TV screen sitting on top of a mobile computer that moves on rollers. In the TV screen, you see the video face of a real physician who may be miles away. There is a camera on the robot that allows the doctor to see what the robot is looking at. There is also a microphone so that the doctor can speak to the patient. The doctor can remotely control the robot via a joystick, interact with patients, monitor drugs, etc. Since annually 5 million patients in the United States are admitted to intensive care units, but only 6,000 physicians are qualified to handle critically ill patients, robots such as this could help to alleviate this crisis in emergency care, with one doctor attending to many patients. In the future, robots like this may become more autonomous, able to navigate on their own and interact with patients.
Japan is one of the world’s leaders in this technology. Japan is spending so much money on robots to alleviate the coming crisis in medical care. In retrospect, it is not surprising that Japan is one of the leading nations in robotics, for several reasons. First, in the Shinto religion, inanimate objects are believed to have spirits in them. Even mechanical ones. In the West, children may scream in terror at robots, especially after seeing so many movies about rampaging killing machines. But to Japanese children, robots are seen as kindred spirits, playful and helpful. In Japan, it is not uncommon to see robot receptionists greet you when you enter department stores. In fact, 30 percent of all commercial robots in the world are in Japan.
Second, Japan is facing a demographic nightmare. Japan has the fastest-aging population. The birthrate has fallen to an astonishing 1.2 children per family, and immigration is negligible. Some demographers have stated that we are watching a train wreck in slow motion: one demographic train (aging population and falling birthrate) will soon collide with another (low immigration rate) in the coming years. (This same train wreck might eventually happen in Europe as well.) This will be felt most acutely in the medical field, where an ASIMO-like nurse may be quite useful. Robots like ASIMO would be ideal for hospital tasks, such as fetching medicines, administering drugs, and monitoring patients twenty-four hours a day.
MODULAR ROBOTS
By midcentury, our world may be full of robots, but we might not even notice them. That is because most robots probably won’t have human form. They might be hidden from view, disguised as snakes, insects, and spiders, performing unpleasant but crucial tasks. These will be modular robots that can change shape depending on the task.
I had a chance to meet one of the pioneers in modular robots, Wei-min Shen of the University of Southern California. His idea is to create small cubical modules that you can interchange like Lego blocks and reassemble at will. He calls them polymorphic robots since they can change shape, geometry, and function. In his laboratory, I could instantly see the difference between his approach and that of Stanford and MIT. On the surface, both those labs resembled a kid’s dream playhouse, with walking, talking robots everywhere you looked. When I visited Stanford’s and MIT’s AI laboratories, I saw a wide variety of robotic “toys” that have chips in them and some intelligence. The workbenches are full of robot airplanes, helicopters, trucks, and insect-shaped robots with chips inside, all moving autonomously. Each robot is a self-contained unit.
Various types of robots: LAGR (top), STAIR (bottom left), and ASIMO (bottom right). In spite of vast increases in computer power, these robots have the intelligence of a cockroach. (photo credit 2.1)
But when you enter the USC lab, you see something quite different. You see boxes of cubical modules, each about 2 inches square, that can join or separate, allowing you to create a variety of animal-like creatures. You can create snakes that slither in a line. Or rings that can roll like a hoop. But then you can twist these cubes or hook them up with Y-shaped joints, so you can create an entirely new set of devices resembling octopi, spiders, dogs, or cats. Think of a smart Lego set, with each block being intelligent and capable of arranging itself in any configuration imaginable.
This would be useful for going past barriers. If a spider-shaped robot was crawling in the sewer system and encountered a wall, it would first find a tiny hole in the wall and then disassemble itself. Each piece would go through the hole, and then the pieces would reassemble themselves on the other side of the wall. In this way, these modular robots would be nearly unstoppable, able to negotiate most obstacles.
These modular robots might be crucial in repairing our decaying infrastructure. In 2007, for example, the Mississippi River bridge in Minneapolis collapsed, killing 13 people and injuring 145, probably because the bridge was aging, overloaded, and had design flaws. There are perhaps hundreds of similar accidents waiting to happen across the country, but it simply costs too much money to monitor every decaying bridge and make repairs. This is where modular robots may come to the rescue, silently checking our bridges, roads, tunnels, pipes, and power stations, and making repairs when necessary. (For example, the bridges into lower Manhattan have suffered greatly due to corrosion, neglect, and lack of repairs. One worker found a 1950s Coke bottle left over from when the bridges were last painted. In fact, one section of the aging Manhattan Bridge came dangerously close to collapse recently and had to be shut down for repairs.)
ROBOT SURGEONS AND COOKS
Robots may be used as surgeons as well as cooks and musicians. For example, one important limitation of surgery is the dexterity and accuracy of the human hand. Surgeons, like all people, become fatigued after many hours and their efficiency drops. Fingers begin to tremble. Robots may solve these problems.
For example, traditional surgery for a heart bypass operation involves opening a foot-long gash in the middle of the chest, which requires general anesthesia. Opening the chest cavity increases the possibility of infection and the length of time for recovery, creates intense pain and discomfort during the healing process, and leaves a disfiguring scar. But the da Vinci robotic system can vastly decrease all these. The da Vinci robot has four robotic arms, one for manipulating a video camera and three for precision surgery. Instead of making a long incision in the chest, it makes only several tiny incisions in the side of the body. There are 800 hospitals in Europe and North and South America that use this system; 48,000 operations were performed in 2006 alone with this robot. Surgery can also be done by remote control over the Internet, so a world-class surgeon in a major city can perform surgery on a patient in an isolated rural area on another continent.
In the future, more advanced versions will be able to perform surgery on microscopic blood vessels, nerve fibers, and tissues by manipulating microscopic scalpels, tweezers, and needles, which is impossible today. In fact, in the future, only rarely will the surgeon slice the skin at all. Noninvasive surgery will become the norm.
Endoscopes (long tubes inserted into the body that can illuminate and cut tissue) will be thinner than thread. Micromachines smaller than the period at the end of this sentence will do much of the mechanical work. (In one episode of the original Star Trek, Doctor McCoy was totally revolted that doctors in the twentieth century had to cut skin.) The day when this is a reality is coming soon.
Medical students in the future will learn to slice up 3-D virtual images of the human body, where each movement of the hand is reproduced by a robot in another room.
The Japanese have also excelled at producing robots that can interact socially with humans. In Nagoya, there is the robot chef that can create a standard fast-food dinner in a few minutes. You simply punch in what you want from a menu and the robot chef produces your meal in front of you. Built by Aisei, an industrial robotics company, this robot can cook noodles in 1 minute and 40 seconds and can serve 80 bowls on a busy day. The robot chef looks very much like ones on the automobile assembly lines in Detroit. You have two large mechanical arms, which are precisely programmed to move in a certain sequence. Instead of screwing and welding metal in a factory, however, these
robotic fingers grab ingredients from a series of bowls containing dressing, meat, flour, sauces, spices, etc. The robotic arms mix and then assemble them into a sandwich, salad, or soup. The Aisei cook looks like a robot, resembling two gigantic hands emerging from the kitchen counter. But other models being planned start to look more human.
Also in Japan, Toyota has created a robot that can play the violin almost as well as any professional. It resembles ASIMO, except that it can grab a violin, sway with the music, and then delicately play complex violin pieces. The sound is amazingly realistic and the robot can make grand gestures like a master musician. Although the music is not yet at the level of a concert violinist, it is good enough to entertain audiences. Of course, in the last century, we have had mechanical piano machines that played tunes inscribed on a large rotating disk. Like these piano machines, the Toyota machine is also programmed. But the difference is that the Toyota machine is deliberately designed to mimic all the positions and postures of a human violinist in the most realistic way.
Also, at Waseda University in Japan, scientists have made a robotic flutist. The robot contains hollow chambers in its chest, like lungs, which blow air over a real flute. It can play quite complex melodies like “The Flight of the Bumblebee.” These robots cannot create new music, we should emphasize, but they can rival a human in their ability to perform music.
The robot chef and robot musician are carefully programmed. They are not autonomous. Although these robots are quite sophisticated compared to the old player pianos, they still operate on the same principles. True robot maids and butlers are still in the distant future. But the descendants of the robot chef and the robot violinist and flutist may one day find themselves embedded in our lives, performing basic functions that were once thought to be exclusively human.
EMOTIONAL ROBOTS
By midcentury, the era of emotional robots may be in full flower.
In the past, writers have fantasized about robots that yearn to become human and have emotions. In Pinocchio, a wooden puppet wished to become a real boy. In the Wizard of Oz, the Tin Man wished for a heart. And in Star Trek: The Next Generation, Data the android tried to master emotions by telling jokes and figuring out what makes us laugh. In fact, in science fiction, it is a recurring theme that although robots may become increasingly intelligent, the essence of emotions will always elude them. Robots may one day become smarter than us, some science fiction writers declare, but they won’t be able to cry.
Actually, that may not be true. Scientists are now understanding the true nature of emotions. First, emotions tell us what is good for us and what is harmful. The vast majority of things in the world are either harmful or not very useful. When we experience the emotion of “like,” we are learning to identify the tiny fraction of things in the environment that are beneficial to us.
In fact, each of our emotions (hate, jealousy, fear, love, etc.) evolved over millions of years to protect us from the dangers of a hostile world and help us to reproduce. Every emotion helps to propagate our genes into the next generation.
The critical role of emotions in our evolution was apparent to neurologist Antonio Damasio of the University of Southern California, who analyzed victims of brain injuries or disease. In some of these patients, the link between the thinking part of their brains (the cerebral cortex) and the emotional center (located deep in the center of the brain, like the amygdala) was cut. These people were perfectly normal, except they had difficulty expressing emotions.
One problem became immediately obvious: they could not make choices. Shopping was a nightmare, since everything had the same value to them, whether it was expensive or cheap, garish or sophisticated. Setting an appointment was almost impossible, since all dates in the future were the same. They seem “to know, but not to feel,” he said.
In other words, one of the chief purposes of emotions is to give us values, so we can decide what is important, what is expensive, what is pretty, and what is precious. Without emotions, everything has the same value, and we become paralyzed by endless decisions, all of which have the same weight. So scientists are now beginning to understand that emotions, far from being a luxury, are essential to intelligence.
For example, when one watches Star Trek and sees Spock and Data performing their jobs supposedly without any emotions, you now realize the flaw immediately. At every turn, Spock and Data have exhibited emotions: they have made a long series of value judgments. They decided that being an officer is important, that it is crucial to perform certain tasks, that the goal of the Federation is a noble one, that human life is precious, etc. So it is an illusion that you can have an officer devoid of emotions.
Emotional robots could also be a matter of life and death. In the future, scientists may be able to create rescue robots—robots that are sent into fires, earthquakes, explosions, etc. They will have to make thousands of value judgments about who and what to save and in what order. Surveying the devastation all around them, they will have to rank the various tasks they face in order of priority.
Emotions are also essential if you view the evolution of the human brain. If you look at the gross anatomical features of the brain, you notice that they can be grouped into three large categories.
First, you have the reptilian brain, found near the base of the skull, which makes up most of the brain of reptiles. Primitive life functions, such as balance, aggression, territoriality, searching for food, etc., are controlled by this part of the brain. (Sometimes, when staring at a snake that is staring back at you, you get a creepy sensation. You wonder, What is the snake thinking about? If this theory is correct, then the snake is not thinking much at all, except whether or not you are lunch.)
When we look at higher organisms, we see that the brain has expanded toward the front of the skull. At the next level, we find the monkey brain, or the limbic system, located in the center of our brain. It includes components like the amygdala, which is involved in processing emotions. Animals that live in groups have an especially well-developed limbic system. Social animals that hunt in groups require a high degree of brainpower devoted to understanding the rules of the pack. Since success in the wilderness depends on cooperating with others, but because these animals cannot talk, it means that these animals must communicate their emotional state via body language, grunts, whines, and gestures.
Finally, we have the front and outer layer of the brain, the cerebral cortex, the layer that defines humanity and governs rational thought. While other animals are dominated by instinct and genetics, humans use the cerebral cortex to reason things out.
If this evolutionary progression is correct, it means that emotions will play a vital role in creating autonomous robots. So far, robots have been created that mimic only the reptilian brain. They can walk, search their surroundings, and pick up objects, but not much more. Social animals, on the other hand, are more intelligent than those with just a reptilian brain. Emotions are required to socialize the animal and for it to master the rules of the pack. So scientists have a long way to go before they can model the limbic system and the cerebral cortex.
Cynthia Breazeal of MIT actually created a robot specifically designed to tackle this problem. The robot is KISMET, with a face that resembles a mischievous elf. On the surface, it appears to be alive, responding to you with facial motions representing emotions. KISMET can duplicate a wide range of emotions by changing its facial expressions. In fact, women who react to this childlike robot often speak to KISMET in “motherese,” what mothers use when talking to babies and children. Although robots like KISMET are designed to mimic emotions, scientists have no illusion that the robot actually feels emotions. In some sense, it is like a tape recorder programmed not to make sounds, but to make facial emotions instead, with no awareness of what it is doing. But the breakthrough with KISMET is that it does not take much programming to create a robot that will mimic humanlike emotions to which humans will respond.
These emotional robots will find their way into our h
omes. They won’t be our confidants, secretaries, or maids, but they will be able to perform rule-based procedures based on heuristics. By midcentury, they may have the intelligence of a dog or cat. Like a pet, they will exhibit an emotional bond with their master, so that they will not be easily discarded. You will not be able to speak to them in colloquial English, but they will understand programmed commands, perhaps hundreds of them. If you tell them to do something that is not already stored in their memory (such as “go fly a kite”), they will simply give you a curious, confused look. (If by midcentury robot dogs and cats can duplicate the full range of animal responses, indistinguishable from real animal behavior, then the question arises whether these robot animals feel or are as intelligent as an ordinary dog or cat.)
Sony experimented with these emotional robots when it manufactured the AIBO (artificial intelligence robot) dog. It was the first toy to realistically respond emotionally to its master, albeit in a primitive way. For example, if you pet the AIBO dog on its back, it would immediately begin to murmur, uttering soothing sounds. It could walk, respond to voice commands, and even learn to a degree. AIBO cannot learn new emotions and emotional responses. (It was discontinued in 2005 due to financial reasons, but it has since created a loyal following who upgrade the computer’s software so AIBO can perform more tasks.) In the future, robotic pets that form an emotional attachment to children may become common.