Up to: What future do we want?


The present level of technology for robotics and communication makes it possible for workers in low wage countries to perform manual tasks remotely in high wage countries. There are plenty of possibilities, but domestic service may be the largest and most straightforward. This alone can be a trillion dollar per year industry if 200 million families in rich countries spend $5,000 per year apiece on remote domestic service, e.g. keeping houses clean. The idea is straightforward.

A standardized robot of adequate visual performance might permit medical examination and some kinds of medical treatment in the home. The patient would have such a robot, presumably the one also used for remote housework, and the doctor or nurse would operate it.

The customer has a mobile robot with arms and a television camera in his home. The worker can see through the camera and move the robot and its arms and hands to perform tasks, e.g. clearing a table, putting dishes in a dishwasher, operating the dish washer, removing the clean dishes and putting them on shelves. The workers can also make beds, pick up after children (and adults), and clean floors.

In low wage countries, the work will be done in buildings with large numbers of terminals and supervisors and translators who speak the language of the customers. Within rich countries, there are people who will benefit from doing such work from home. In fact, this may be the place to start, because communication will be easier.

As with every other export of work from rich countries to poor countries, this will develop the economies of the poor countries while raising the standard of living at both ends. Eventually, the wages in the poor countries will rise to the point where the business is impractical, but this is likely to take a very long time.

Here are some relevant facts.

  1. The technology and cost of communication are adequate to transmit television images between any two places in the world of the quality needed to control robots in real time. The delays associated with satellite transmission may make tasks difficult, but fiber optic cables are now the main means of point-to-point transmission and have adequate speed.
  2. Present mechanical robot technology is adequate for doing housework, but robots combining the technologies will probably require further development.
  3. The present state of automatic robotics and AI is not up to doing many operations now done by people that cost much less in poor countries. Examples include barbering and domestic service. Current robots can carry trays around and not a lot more. The problem is the programming rather than the hardware.
  4. A current active area of research is remote surgery. In this case, the object is to magnify the efforts of an expert by allowing him to do surgery without travelling to the patient. There are some extra possibilities; the remote tool need not correspond exactly to the surgeon's hands. Professor Jon Bowersox at Stanford has a project on telesurgery supported by DARPA.
  5. Millions of people from poor countries take jobs in rich countries. Some migrate legally as permanent residents. Others come under arrangements that are nominally temporary, i.e. they are expected to work for a while and eventually return home. There is substantial illegal migration. Employers in rich countries can get people from poor countries to work for wages which are lower than those required by citizens but are higher than the workers can get in their home countries. Besides business employers, there are large numbers of private individuals employing foreign domestic workers.

  6. Every rich country puts restrictions on the number of immigrants from poor countries. Immigrants can become burdens on the welfare system, and if they bring their families, the families become burdens.

  7. The ratio of wages in many occupations between rich and poor countries is large enough to sustain a lot of overhead.

  8. Unless something surprising happens, China and India will sustain the low wage business for a very long time.

More on the Technology of Teleservice

The problem has several aspects.
Satellite communication may impose excessive control delays for fine work, especially if two satellites are involved. This is because communication satellites in geostationary orbits are 22,000 miles from the earth. With two satellites that is 88,000 miles, and a back and forth is 176,000 miles, which is almost a second at the speed of light. Fortunately, fiber optic cables are becoming the main method of intecontinental communication. For them, the delays are much smaller.

TV transmission and compression
Most likely the costs are low enough so that a full two-way TV channel is quite cheap. Otherwise, there are many possible bandwidth saving schemes, and these will depend on the application. For example, much of the scene remains unchanged as the remotely controlled robot moves. Evidently much of the relative positions of parts of the customer's head remains unchanged during barbering. It has recently been announced that TV can be transmitted over a 28.8kb modem.

Robots suitable in their mechanical properties for barbering and house cleaning may not exist yet, but they seem to be within the state of the art of mechanical engineering. To be widely usable, a house cleaning robot should not weigh much more than a person and should be controllable to climb stairs.

Force feedback
Some degree of force feedback (haptic feedback) to the workers may be needed. Experiments in the Stanford Robotics Laboratory (Professor Oussama Khatib's work) offer evidence for the need and means for its accomplishment.
Program assistance
AI seems to be slow in developing robots that can do these jobs all by themselves. However, it is good enough to do parts of them and thereby enhance the efforts of the human worker. Gradually, more and more of the control will be robotic.


Consider the effects of remote service on two occupations - barbering and domestic service.


A barber in Bangladesh cuts just as much hair per day as a barber in the U.S. - maybe more, since his hours may be longer. Very likely he is just as skillful. Economists give barbering as an example of an occupation that cannot be exported and therefore commands wages determined by the productivity of other occupations in the same country. Economists were mistaken in believing that barbering could not be exported.

In 1987 American barber and beauty shops had 390,000 employees. The large scale use of foreign based teleservice might reduce this to less than half, perhaps to less than 1/10. It would depend on the development of very specialized services and on the need for supervisers of the foreign operations. Barbering is a rather stable profession, and barbers in the developed world would resist being displaced. When linotype operators were displaced, there were employers who could be induced to buy out the jobs. This is not the case for barbers. Probably the Government would make it easy for them in some way, perhaps by a temporary tax on remote barbering. For these social reasons, probably the displacement of barbers will be slow or long delayed.

Domestic Service

Domestic service is less problematical socially than barbering. Everyone gets haircuts, so the main effect of remoter barbering will be to displace current barbers. The largest effect of remote domestic service would be that people who currently can't afford domestic service would hire remote workers.

The overall effect of teleservice will be to gradually equalize world working conditions and living standards, mainly by bring up the standards in the less developed world. This is the same effect that moving manufacturing to the underdeveloped world is having. Eventually, Chinese barbers and domestic workers will be paid as much as Americans, but this will only happen through development of the Chinese economy in all its sectors. Because of the enormous populations of China and the Indian subcontinent, they will remain reservoirs of inexpensive labor for a long time.

How will it look in a developing country?

(I am indebted to Michael Gaisford for his 1998 STS160 essay Teleservice: From the Other Side. It convinced me that this question needs more discussion.) Here are some considerations:

The simplest case is that teleservice is not the pioneer industry producing for export. There will be some foreign owned facilities and some native owned facilities. The governments of developing countries almost always insist on getting a substantial piece of the action for their own citizens even if they have to require that government officials or their relatives are part owners. In this case, there will already be an industrial labor market around the larger cities and substantial migration into the industrial area from rural areas.

The local end of the teleservice companies will then set up offices where the teleservice will be performed. There will be supervisors, translators and technicians to maintain the controllers and the communications equipment. There will also be a training school, perhaps complete with a "typical home" in the U.S., Japan or South Korea.

The workers will already know about the West from movies and TV, most likely having an image more glamorous than the average reality. It will be a come down for someone who imagines working in the mansion of a movie star to working in the home of carpenter or an accountant. They'll get used to it quickly enough.

It will be mutually advantageous if the work is done repeatedly in the same home, and, if language permits, a personal relation may develop between the employer and the employee. If communication facilities to homes permit, it may become advantageous for some of the workers to become free lance domestics, relying on greater skills and better personal relations to command higher pay and to cut out the middle man. Sometimes the operators of the service companies will resist this, but they will compete with each other, and there will be a reduction in supervisory expenses when the worker is on his own.

We can be sure that the industry will take surprising forms that won't be anticipated in advance.

We can also imagine teleservice starting in more backward areas, but this is unlikely to happen until (1) the pump is primed by governments in order to improve their economies or (2) the supply of cheap labor in the more advanced backward countries runs low. It will be quite expensive for teleservice companies to set up facilities where there isn't already a usable electricity supply and telephone facilities. However, these facilities are important for manufacturing of the conventional kind in low wage countries.

Matthew Johnson suggests that driving cars is a feasible and useful form of teleservice. I agree.

Up to: What future do we want?

Send comments to jmc@cs.stanford.edu

The number of hits on this page since 1995 Jan 21.