Sunday, August 13, 2000		Article

FOR those of us who love both science-fiction and sci-fi movies, computers that have feelings and robots that look and act like humans are staple diet. Isaac Asimov, the Grand Old Man of science fiction painstakingly created an intricate world in which robots worked for and with humans. The robot nursemaid which loved its charge to distraction, the robot which wanted to be alive even

though that meant it would be mortal, the robot which malfunctioned and told tales to cause trouble in a robot research factory and the

android which successfully masqueraded as a human in a world where robots were hated enough to be destroyed at sight. Asimov even laid down the three laws of robotics. The first law of robotics states that a robot may not harm a human, either through action or inaction. The second law enjoins the robot to always obey humans unless the command comes in conflict with the first law. One consequence of this law is that potential murderers cannot order their robot to kill a human. The third law states that the robots must protect themselves but not at the cost of violating the first law. Of course, as in the human world, there are devious means of circumventing these laws, on the basis of which Detective Elijah Bailey and the android Danieel Olivaw solved many a mystery.

Unfortunately, androids and robots that look like humans and ‘think’ creatively, are as distant as settlements on the Moon. Even the simple act of walking on two ‘legs’ is still a mere probability would-be androids. The most stable ‘walker’ so far has been a six-legged device. This seems to lend credence to the view held by the French philosopher-mathematician Rene Descartes, who declared in 1637 that it would never be possible to create a machine that thinks as humans do. His belief was contradicted in 1950, by British mathematician and computer pioneer Alan Turing who insisted that one day there would be a machine which would duplicate human intelligence to such a degree that if a machine and a human were asked the same question, the questioner would be unable to distinguish between the answers given by the machine and the human.

Nearly 50 years later, no machine has come up to the standard set by Alan Turing although by the early 1990’s computers could ‘make decisions by interpreting data and

selecting amongst alternatives. For example, a chess-playing computer has to be fed all conceivable permutations and combinations of the pawns before it can play with

the likes of Vishwanath Anand. Of course, this involves the tedious process of extracting even the most infinitesimal iota of knowledge from expert sources and then feeding it into the computer. But the effort is worth it because the human brain simply cannot match the analytical and computing speed of a computer, the fastest computer being able to perform approximately 10 billion calculations in one second. These computers, which reflect the knowledge of myriad experts and make selections on the basis of the data they possess are known as ‘expert systems’. Two expert systems in use are Internet, a computer system which accurately diagnoses 550 diseases and disorders and Prospector, which aids geologists in

detecting new sources of mineral deposits. However, geologists have to feed information from maps, surveys etc to Prospector before it can indicate a possible mineral

deposit while Internist needs to be fed each detail of the patient’s history before it can make a correct diagnosis. A far cry from the robot doctors of Star Wars which repair broken limbs and provide Luke Skywalker with a robotic replacement for his lost arm.

The origin of the usage robot itself defines the characteristics of an ideal robot. In 1921, a Czech play written by Karel Capek called R.U.R. about mechanical beings manufactured to be slaves rebelling against their human masters, gained immense popularity. It was translated into English in 1923. In Czech, the word robot literally means worker. So one can say a functional robot is one which can replace a human at performing a job it is programmed to do with no one knowing the difference. To date, the robots that have been developed have been able to perform only specific functions and have not done away with the necessity of human presence. Since the 1960s, robots have been used in assembly lines to perform tasks like welding, painting and inspection. Others manufacture integrated circuits. With advanced computerisation, they are being used for more delicate tasks. For instance, robotic arms equipped with surgical tools can assist surgeons in performing delicate operations. They are also being used in scientific research, in military programs, as educational tools and to aid people who have disabilities. The latest is a firebot, a robot that takes over the responsibility of fire-prevention from humans. Programmed to respond to the sound of a fire-alarm, it comes armed with a fire-extinguisher. It has the ability to move from room to room, searching for flames through sensors and can put out these flames with the extinguisher. It even takes the bother of calling the fire-brigade away from humans. This is probably the first robot that has practical appeal to house-holders.

In scientific research, robots have helped push back the frontiers of study as they are not limited by the frailty of the human body. They enter volcanic craters and plunge into underwater canyons. Since

they are immune to the physiological consequences of exposure, they can handle radioactive and other hazardous materials. They have substituted for humans in

studies carried out on Mars and Venus. These highly automated systems called probes have also studied the soil of the moon. Robots have even, when the need arose, been used as a secure ‘space platform’ for astronauts to work from outside their spaceships. Those of us who have seen the film Titanic have been a robotic camera in action. This robot not only captures images of the interior of the sunken ship but also picks up objects as delicate as a woman’s hair ornament and as heavy as a safe. But it cannot decide independently whether an object is of possible interest or not. It will pick up all the objects it is instructed to, the trash along with objects of interest. Briefly, it lacks reasoning.

The robots developed so far cannot work independently but respond to the instructions of carefully-designed, complex computer programmes. At the most basic level, we can call them machines run by computers. They cannot learn new tasks and have to be reprogrammed to respond to a new set of instructions. They cannot react to any change in their environment nor can they take independent decisions.

Some scientists working on artificial intelligence have concluded that the best way to design a reasoning machine is to create a design based on the human brain. The functioning of the human brain is essentially carried out by a network of nerve cells or neurons. Scientists reason that if chips, the DNA or the building blocks of a computer, can be made to behave like neurons, a network of these chips will simulate the human brain. By the early 1990s, such a silicon chip had been developed and in December 1996, scientists at the Argonne National Laboratory in Illinois claimed to have programmed a computer that could simulate human reasoning. This computer reasoned out the solution to a complex mathematical problem independently, following mathematical rules rather than a specific set of instructions. If scientists can steer the development of the reasoning computer into a direction which will lead to the development of machines which possess vision and touch sensors and share knowledge not just with humans but with each other, we might see machines pass Turing’s test in our lifetime.

This fascination with mechanical devices created only to ease human life has existed since the 8th century BC when Homer wrote the Iliad, in which he described "handmaids of gold resembling living young damsels". But the author who seems to have the most accurate vision of the development of computerised devices in the 20th century is Arthur C. Clarke. His fictional world of space travel in the 21st century does not have androids. Instead, it has a computer called HAL9000 which is so perceptive that it suffers from the extreme emotions that some acutely sensitive humans possess. In a even more distant future, a reasoning, mechanical, independent being, capable of replicating itself ponders on its origin the way humans have since intelligence dawned. It rejects the claim of the inhabitants of a planet called Earth that they are its creators because, it

argues, how can beings of an inferior intelligence have possibly created the super-being that it is. Can the makers of Internist and Prospector predict such acumen in future generations of their creations?