The idea of a robot, a purely mechanical device that can perform all manner of tasks normally associated with humans, has been around for a long time. The word itself was used in a play back in 1920, Karel Capek's "R.U.R," standing for Rossum's Universal Robots. Given the vast progress that we have made during the intervening 80 years in manufacturing, electronics, miniaturization, automation, and most especially in computers, you have to ask, what happened? Why are there no robots?

One answer is, there are. Today's factories are full of sophisticated devices that separate defective from acceptable components, move large objects around, and perform precise work on assembly lines. These are all robots. However, they are special-purpose machines, usually working at a fixed location and performing a limited number of operations. They do these things very well. By contrast, the mobile robots that have been built so far are slow-moving, clumsy morons, trundling their way with difficulty across a simplified environment to pick up colored blocks with less skill and accuracy than the average two-year-old. They are a far cry from the robots of fiction, which do general work around the house and handle everything from washing-up to child-minding.

Where are we going wrong? Why have we not been able to build general-purpose robots?

The main problem seems to be the distorted way that we humans see the world. We have these big and complex brains that set us apart from every other animal, and as a result we tend to overemphasize the importance of logical thought. At the same time, until recently we have underestimated the difficulty of all the functions that we share with animals. Those functions include the easy recognition and understanding of environment, grasping of objects and effortless locomotion across difficult terrain.

But seeing and grasping and walking have a billion years of development effort behind them. We do them well, not because they are simple, but because evolution weeded out anything that found them difficult. We don't have to think about seeing. Logical thought, on the other hand, has been around for no more than a million years. No wonder we still have trouble doing it, and are so impressed by it. We are proud of our ability, but a fully evolved thinking creature would find logical thought as effortless as taking a drink of water. Moreover, most organisms in the world get along very well without any ability to think at all, in the sense that we use the word. Ask a worm, or even a dog, what's on its mind today.

Recognizing the nature of the difficulty does not solve the problem, but it does allow us to place emphasis on the appropriate area. For many years, the "difficult" part of making a robot was assumed to be the logical operations. This led us to develop computer programs that play a near-perfect game of checkers and a powerful game of chess. The computer hardware and software combination known as Deep Blue beat world-champion Gary Kasparov in a series of chess games in 1997. During that contest, the computer making subtle and far-reaching strategy decisions in the game was as helpless as a baby when it came to picking up a chess piece and executing a move. Those functions were performed by Deep Blue's human handlers.

So when will we have a "real" robot, one able to perform useful tasks in the relatively complicated environment of the average home?

The answer is: when development from two directions meets. First, we must have the "top-level" capability, usually referred to as Artificial Intelligence, or just AI, that seeks to copy human thought processes as they exist today. AI, after a promising start in the 1960s, stumbled and slowed. One problem is that we don't know exactly what human thought processes are. As one pioneer in the field of AI pointed out, the easy part is modeling the activities of the conscious mind. The hard part is the unconscious mind, inaccessible to us and difficult to define and imitate. Second, we must have "low-level" capabilities that solve the basic problems of perception and mobility, without introducing the idea of thought at all. This calls for an "evolutionary" approach, building systems that incrementally model the behavior of animals with complex nervous systems. We know that this can be done, because it happened once already, in Nature. However, we hope to beat Nature's implementation schedule of a few billion years.

It should be clear by now that the terms "high-level" and "low-level" represent a peculiarly human bias, and in no way reflect the importance or difficulty of the two types of functions. However, when the high-level and low-level capabilities converge and finally meet, in what has been termed the "golden spike of robotics," the result will be a reasoning computer program able to perceive, decide, and move at near-human levels; in other words, a robot as they have been featured in fiction for so many years.

That suggests the next question: When? The leaders in artificial intelligence and robotics say we will have "real" robots within 40 years. They are perhaps not entirely impartial, but if we do not accept the estimates of leaders in a field, whom do we believe?

We still must ask one more question: When we have robots able to reason at the level of humans, and perform the same physical and mental functions as humans, what will come next?

The optimists see a wonderful new partnership, with humans and the machines that we have created moving together into a future where human manual labor is unknown, while mental activities become a splendid joint endeavor.

The pessimists point out that computers are only half a century old. In another century or two they may be able to design improved versions of themselves. At that point humans will have served their evolutionary purpose, as a transitional state on the way to a higher life form. We can bow out, while computers and their descendants inherit the universe. With luck, maybe a few humans will be kept around. Let's say, as historical curiosities. Let's not call them pets

"Borderlands of Science" is syndicated by: