Click for next page ( 155


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 154
Computers and Business THEODORE J. GORDO.N Computers and automation are so irrevocably entwined with business that it is hard to imagine what business would be like without them. These technologies computers and automationare ubiquitous, ex- pected, and necessary, appearing in almost every facet of business enterprise: from recruitment through layoff, from raw material ordering through the manufacturing of products, from identifying sales prospects to order entry and delivery, from competitive analysis to strategic optimization, and from innovation to design applications of computers and automahon. They permeate business life, and in doing so have changed it for all time. Yet there is more to come, not only with respect to business applications of computers and automahon, but perhaps more importantly with respect to their impact on business itself and on the people who run it. This is a vast and literally boundless topic, so some structure is necessary if we are to discern even the highlights of prospective change. A three-dimensional space serves as the organizing principle for this paper: Business functions comprise one axis; on another axis lie the technologies; and on the third, the impacts. For convenience, I have divided business functions into management, manufactunng, selling, planning, training, and professional support. For technology I have defined three major facets: computer hardware and software; programmable automation, which includes robotics; and telecommun- ications. And in the third dimension, the impacts fall on three elements of business: people, those who define and execute the intent of business; 154

OCR for page 154
COMPUTERS AND BUSINESS 155 structure, the organization of the business ente~pnse; and self-concept, the view of what business is about and is capable of attaining. This structure leads to a set of interesting questions. For example, How will future developments affect people in management? For this discussion I have selected only a few of these intersections, on the basis of their significance, the number of people affected, and the likelihood of the developments in the next 10 to 15 years. This is certainly not a complete set, and it relies heavily on judgments about what is possible and about the complex processes of acceptance and response to new technologies. The changes we have seen to date are staggenugly significant. What is coming, however, not only extends the trends of the past but includes massive changes that, in the aggregate, will define the very nature of business and the relationships between those who serve and are served by it. My approach win be to set the stage with a discussion of some characteristics of emerging technologies and then to describe some of the changes that seem likely at four intersections in our three- dimensional (business function/technology/impact) space. The inter- sections and their respective changes are as follows: At the intersection of training, computers, and people, the discussion centers on simulation as an aid to training; At the intersection of manufacturing, automation, and people, on factory automation and robotics; At the intersection of selling, telecommunications, and structure, on retail electronic funds transfer (EFT); and At the intersection of management, computer, and self-concept, on modeling in management decision making. SOME CHARACTERISTICS OF EMERGING TECHNOLOGIES First, the stage setting. There are three principal hardware trends that characterize electronic hardware today: reduction in cost, im- provements in reliability, and increases in packing density (the number of components that can be packed into a given volume). Each of these characteristics of the technologycost, reliability, and density has been changing by approximately a factor of 100 each decade since 1960. Studies by The Futures Group indicate that these trends can continue for another two decades or so. As limits are reached during this period, new technologies will offer potential for further break- throughs. For example, photolithography (the technology required for printing microcircuits on silicon) is limited at present by the distance between lines that can be drawn optically. This, in turn, is fixed by the wavelength of light. Once this limit is reached, conventional

OCR for page 154
156 THEODORE; ]. GORDON photolithography impedes further progress toward miniaturization. However, just behind this conventional technology lies the possibility of using shorter-wavelength energy in these processes; for example, electron beam or X-ray imaging. In addition to these more or less continuous trends of improved reliability, reduced cost, and increased packing density, it is worth noting two other developments of significance; these are discontinuities that can significantly affect the application of electronics in the future. First, the same techniques that are being used to produce very large scale integrated circuits are also being applied to the manufacture of small mechanical devices. For example, a mass spectrometer, a device for determining the constituent elements of gases and other fluids, has been "printed" on a chip. This is more than simply printing the electronics on a chip, as is commonplace in very large scale integrated circuitry. Rather, the whole machinevalves and all is part of the printed apparatus.' An example of the future use of Microsystems in industry is the potential for process instrumentation that floats with a stream of chemicals and telemeters process control data (rather than being point-fixed on the well of a pipe). Another industrial example is temperature and pressure instrumentation built into a grinding wheel or cutting tool to control a feedback system that optimizes metal removal rates, improves precision of manufacture, and extends tool life. In the office these m~cromechanical devices can be useful in constructing extremely small microphones for dictation, telephones, or security systems (also, micro eavesdropping bugs), or feedback instruments for chairs, printers, and personal, local air conditioning. With this technology, buildings can be instrumented to detect incipient mechanical failures, and quiescent manufactured products can tele- meter their state of functioning or readiness in response to an external radio trigger signal. In short, micromechanics not only will permit replication of macromachines on a tiny scale but will stimulate the innovation of entirely new applications that benefit from small size, low pnce, dispersion, and decentralization. The second development is the coming of ageprobably within the next decadeof artificial intelligence (AI), the simulation of human intelligence by computers. Artificial intelligence requires the ability to sense, operate on sensed information, draw inferences from observa- tions, and perform adaptively in view of these inferences and changing circumstances. AI programs have two general attributes: search and knowledge. Search comes from "defining a space of possibilities large enough to contain the sought-for solution," and "knowledge is nec- essary to guide the search through the space."2

OCR for page 154
COMPUTERS AND BUSINESS 157 The initial entries in this field are expert systems that capture pragmatic if-then rules of analysis followed by human experts in a given field. Examples of expert systems that are functional or in development abound. Medical diagnostics, personal financial services, geological exploration, legal strategy, and software design are some examples. In these fields and in others yet to come, computers will produce practical and functional answers to real problems better than answers that could be produced by a random sample of professionals. When expert systems can learn from expenence, the decision rules incorporated in the knowledge portion of the program can be much more extensive, and a transition will have occurred from programs that merely emulate the behavior of experts to creative artificial intelligence routines. These routines will arrive at answers better than those that might be created by most human experts. This will not happen tomorrow, but it is reasonable to expect self-learning systems to be in operation within 15 years. This image of the growth of electronics and its applications and spin- offs depends, of course, on the market. Given demand, the capability of electronics and all of its derivatives grows; without demand, nascent applications wither. Business fosters demand both by offering new products to consumers and by becoming a consumer itself. In either mode business touches and is touched by these technologies and thus is changed, not only in how it does business, but in what it believes business to be. Let us turn now to some intersections in the three- dimensional space discussed earlier. INTERSECTION: TRAINING, COMPUTERS, PEOPLE Simulation as an Aid to Training Take as a starting point high-density TV (in the more distant future, perhaps holographic TV) driven by rapidly accessed videodisks, excellent and sensitive computer simulation programs, and much more effective input-output systems. Put them together and imagine a worker- training system 10 to 15 years hence. The key here is softwaresuperb simulation techniques that permit the creation of accurate environments that stress the student and promote learning. With such software, the emphasis in education switches from teaching to learning. Pilot training serves as a current example. In the future, decreasing equipment costs, better software, and realistic input-output systems mean that appli- cations will be far less monumental and will certainly be applied in business for training of production workers, managers, salespeople,

OCR for page 154
158 THEODORE J. GORDON repairmen, and anyone likely to benefit from the stress of practice. Simulation in the context used here means that the user determines the plot of an unfolding story through his or her decisions. This is computer game playing carried to its logical end. In schools you can be with Napoleon or serve as a lieutenant to Washington. In leisure at home, TV becomes active rather than passiv~here you can be with Cleopatra or J. R. Ewing. In the workplace you can direct field operations to put out oil well fires, or learn, as a potential manager, how it feels to deal with labor grievances and the chess game of a stnke. Consequences Personnel training and education will be substantially changed, with competitive advantage falling to the best corporate simulators. Imagine the new case study approach in management education: "Have you played Continental Illinois yet?" Grading of personnel in corporations may be according to their successes or failures in simulations. Imag- in~in awed tones "He made Continental survive." Perhaps bore- dom will become a problem when the game environment makes the adrenaline flow more easily than reality does. New products and markets will be built around interactive on-the- job expenence. When the simulation tools are very good they will also serve as decision-aiding tools. For the repairman not quite sure about which wire to connect, a quick run on the simulator will show the consequences of connecting the wire a working machine or a blown fuse. The analogous situation for management dilemmas is obvious. INTERSECTION: MANUFACTURING, AUTOMATION, PEOPLE Factory Automation and Robotics When computers were first introduced, there was a great deal of concern that unemployment would result. In fact this was not the casewherever computers were used, more jobs were created. Com- mon wisdom holds that this situation will always continue, but it might not be so. Factory automation is likely to advance so far beyond current capabilities that the net effect of introducing such new tech- nologies may be to improve total output with less labor required in both a relative and an absolute sense. I will explore the potential for such technology-induced unemployment in this section. Artificial intelligence will allow machines to perform cognitive

OCR for page 154
COMPUTERS AND BUSINESS TABLE 1 Forecasts of Robot Technology 159 Sate of the Art Feature 1983 2000 Accuracy of manipulation (Electnc) 0.02 inches 0.001 inches (Hydraulic) 0.2 inches 0.020 inches Repeatability of placement (Electnc) 0.005 inches 0.001 inches (Hydraulic) 0.050 inches 0.010 inches Mean iune between failures 1,000 hours 5,000 hours Fault detection and repair Mostly human Mostly self-check Speed for standard pattern 4 seconds 1 second Programming External External and self- taught Sensing visual Silhouette 3-dimensional Memory capacity and type Magnetic media Vastly expanded magnetic and op- tical media Infonnaiion processing Sequential archi- Parallel architecture lecture SOURCE: The Futures Group. functions; robotics will move from specialized to general-purpose applications. Recent studies at The Futures Group have resulted in projections of robot technology that illustrate the enormous potential for this field. As Table 1 shows, robotic accuracy, repeatability, mean tune between failures, time to repair, and speed are expected to increase significantlymore than an order of magnitude in most cases in the next 15 years. Within the next decade or so we also can expect to see very simple means for programing robots and, with the advent of artificial intelligence, robots that learn through experience For example, a robot could be adaptively programmed to change its positioning or sequence in order to ~nirnTn~e rejection rates. A distinguishing feature of robots is their versatility their ability to be used in a multiplicity of applications. In the future, robots will become more general-purpose, in the sense that their implements can be utilized in a variety of jobs without much cost penalty. Vision and sensing will improve to make three~imensional perceptions common- place. The number of robots to be employed in the future is not certain by any means, but large increases seem likely. The number of robots in manufacturing quadrupled between 1979 and 1981.3 Forces encouraging growth include

OCR for page 154
160 . TH~ODO~ J. GORDON - Improvements in the technology itself, which increase the number of applications possible with these machines; Diminishing costs for given robotic capability as a result of learning-curve improvements; Increasing cost of human labor; and Growing sophistication on the part of management, facilitating the switch to robotics. On the negative side, factors that limit the speed of diffusion of this technology include The size of the required investment, Institutional inertia that slows the adoption of automated technologies, and Obsolescence of current equipment. Now the question is: Will progress in robotics and factory automation in general create jobs or eliminate them? The answer is, of course, it will do both. At constant levels of output it will eliminate jobs, because the robots will perform jobs that human workers currently perform, and automation, properly applied, will improve productivity through increasing output per man-hour. Some people argue that as automation progresses more people will be required to produce the machines and that as people are freed from dull, repetitive, boring, and sometimes dangerous activities, unemployment will not diminish but, in Parkin- sonian response, the scope of work and perhaps its quality will increase to occupy the new capacity for work. While this has generally been the case in the past, robotics and the new wave of automation have some new attributes. They promise to be very good, very cheap, and unt~nng. They can yield manufacturing quality higher than that pro- duced by their human counterparts. Instructing these machines (pros gramming) will be efficient and, through AI, they will be adaptive; mu an V; ~5 one smug programming may be, the machines will progress and learn to do even better The machines will c-if Air t"~^ ~~ __~: ~ _ ^~ _ ~ ~ . ~ . . ~^ a_ ~ ~~ TV ~~ ~ 1 a_ _' __ ~ . . ~ . - _ learn OI impending internal failures through "introspection," and- when necessaryself-repair, to a small extent initially and to a major extent later. More than that, when programmed to do so, they can self-replicate. Just how many jobs will be displaced by continuing factory auto- mation and robotics? The situation is summarized in Table 2. We have created a scenario with several critical assumptions. The U.S. De- partment of Labor expects the labor force to grow from its present level of 110 million to about 134 million by the year 2000.4 We assume that real productivity grows at about 1.5 percent per year as a result

OCR for page 154
161 C) o s: ~ o C~ ~ ~ _ & s:~- La ~ O 3 CC U: - ~_ O ~ ~ ~ _ ._ O O O - ~0 ~ C~ Ct O .o & & ~ ~ O _ ~ O .> ~ O _ O O Pt O ~ s: C~ C~ O ~ O ~ O S Z Pt _ o g X - . o . - S~ C) o ,1 m ~_ 0 ~ ;^ ~ _ ." 0 0 ~ 3 ~ _ & O oo _ ~ ._ _ C,_, _ ~ O ~ U) .3 _ ~ & 3 O o Z ~ O C; C) S o ~ C) ~ U~ O CC O o o ~: O ~ ._ - C) - & o o~ r~ s~ s~ r~ - oo O O _ _ oo o x - o - - - ~ o o o o o ~ o ~ o o o - ~ r~ O O O Q O ~ o _ ~ ~ * O ~D ~ O ~ 0\ ~ ~ ~C O ~ O ~ O O C~ ~ X ~ oo ~ t ~ X ~ oo ~ 0\ ~ O ~ ~ X ~ . . . . . . . . . . . . . . . . . . . O ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ X oo ~ ~ O Ch~ ~ ~ d ) ~) CP~ ~ ~C 00 } ~) ~ ~ ~] O 0\ oo oo X oo ~ o _ ~ ~ ~ ~ _ O ~ ~ ~ ~ ~ ~ ~ oo ~ ~ (~} ~ t<) ~ oo Ct _ _ _ ~ ~ ~ ~ ~ ~ oo ~ O ~ ~ ~ ~ ~ ~ oo C~ O oo oo oo X X X oo oo C~ C~ ~ ~ ~ ~ ~ C~ ~ ~ O ~ ~ ~ ~ c~ ~ ~ ~ ~ ~ c~ cr, ox cr~ ~ ~ ~ cr~ 0 C C) o C~ V) C: o ._ 3 ~ o co ~ . o ~ _ C ~ 0 .^ ~ C) _ CJ Ct ~ C) 0 0 0 ~o _ ~ .O s C) _ _ 3 Ct o oa C~ C~ 2 6, C) ' .. CO ~ : o~ ._ ,_ o. .' _ cn .0 o <, .. ~ O ~ ~ * o V U, C) - s . . o C~

OCR for page 154
162 THE ODORS; J. GORDON of automation and that real Gross National Product (GNP) grows at 2.7 percent per year. Furthermore, the number of robots is assumed to grow from a currently installed base of about 5,000 to 500,000 by the turn of the century. We suppose, further, that the effectiveness of each robot also grows; today a robot is equivalent to about two persons, and we have assumed that by the turn of the century a robot can replace five workers. With these assumptions, as Table 2 shows, the contribution of robots is relatively minor. Only about 2 percent of the labor force expected in the year 2000 will be displaced by robotics. While the picture presented in Table 2 is a homogeneous represen- tation of the labor force as a whole, certain industries will be more affected by robotics than others. In general, these are industries in which mechanization of production yields lower cost, higher quality, diminished production time, improved efficiency, or improved worker safety. In these industnes, the impact of robotics on job displacement will be considerable. For example, the production of passenger auto- mobiles has involved a labor force of about 270,000 over the last seven years. On average, this labor force produced about 30 automobiles per employee, while production during this interval varied from 6.2 million to 9.2 million units per year. Now, assume that the number of individuals available for passenger automobile production grows at the same rate as the labor force as a whole. The employees required, however, are affected by level of production of automobiles and by improvements in productivity resulting from factory automation and the introduction of robots. If we assume (1) that production grows at 3 percent a year (so that by the year 2000 more than 10 million units are manufactured in the United States), (2) that productivity grows at 1.5 percent per year (as previously assumed), and (3) that the number of robots used by the industry grows from 5,000 in 1985 to 25,000 in the year 2000, less than half of those who might ordinarily have been assumed to be available for employment in this industry will be required. A major public policy concern, of course, is that those displaced may be the least able to find new employment. Whether this is a barrier to the spread of factory automation depends on many factors, such as the job security provisions of labor contracts in affected industries and the state of the economy. My guess, however, is that robotics and programmable automation will spread rapidly as the return on invest- ment in such systems grows. Consequences There is likely to be a Wowing emphasis on the job security issue and retraining in industries likely to experience displacement. Also,

OCR for page 154
COMPUTED kD BUSINESS 163 there may be the return to the United States of some jobs previously located outside the country to take advantage of low-cost automated production. Product quality will improve and a new class of employ- ment such as the blue-collar programmer will emerge. Factory automation and robotics will constitute a new technological frontier on which the battles for international markets will be fought, since the products these technologies yield will be of lower price, higher quality, and more predictable performance. INTERSECTION: SELLING, TELECOMMUNICATIONS, STRUCTURE Retail Electronic Funds Transfer (EFT) There is an unusual and, to some extent, unexpected confluence of technological and consumer trends that may affect the retailing envi- ronment in the immediate future. I believe that automated debit purchasing may come on the scene faster than many people expect. Here are some of the factors that lead to this position: Automatic teller machines (ATMs) have spread (47,000 units in place in 1983) and have become much more widely accepted by consumers.5 The principal use of ATMs has been to withdraw cash from personal accounts. Many retail establishments are considering or are installing ATMs on their premise~small bank branches in order to provide a means for their customers to obtain cash in their stores. The communication networks required to support the ATMs and the software necessary to properly debit accounts exist and are proven. It takes only a small step of imagination to move the ATM into the cash register so that at time of purchase consumers can simply insert the ATM card into the proper slot, punch in their personal identification number, and be charged for the purchase directly, just as if they had gone to the ATM and withdrawn cash. In this view, there will be no such thing as a special debit card. The bank-issued ATM card takes its place and the era of electronic fund transfers at the retail level will have come on the scene smoothly, with minimum fanfare, and with relatively high consumer acceptance. More than just a convenience to shoppers, this is a crucial and catalytic step to a society that, while not cashless, certainly functions with much less cash. An important by-product of this development for business will be the availability of really exquisite information about who buys what, where, and when the basis for a potential revolution in marketing research. Where will this trend surface first? Probably in supermarkets

OCR for page 154
164 THEODORE; J. GORDON and gas stations. It then spreads in the retail environment wherever population density and purchase traffic are high enough to warrant the investment. Then it may spread to vending machines of this new era. These vending machines will be operated by coins as well as by ATM cards. What better way could be found to reduce store theft? The image for these vending machines will be upscale, and I would guess that these machines will typically carry much higher-value items than do machines of today. One feature of these machines that will make them unique is their ability to be easily programmed to offer a wide variety of merchandise (sweaters, socks, small electric appliances, stationery). In this way the machines can be generic, and the price can be set for whatever merchandise the retailer wants to sell. Consequences The rate of growth in the number of checks written and processed will be greatly reduced, as will the amount of mail. There will be a major new role for secure telecommunications networks, and today's credit cards will be transformed into their debit-credit equivalent. Once the networks have linked consumers with their accounts at points of sale, whole markets can go electronic, allowing buyers and sellers to meet electronically, and bid and auction until deals are made remotely. It seems to me that this kind of market, based on network connections, could be very well suited to real estate, tax shelters, and any other high-value transaction in which "feeling the merchandise" is not essential. INTERSECTION: MANAGEMENT, COMPUTER, SELF-CONCEPT Modeling in Management Decision Making The goals of a business are, by and large, determined by what management believes is feasible at acceptable levels of investment and risk. These perceptions are, in turn, informed by data available to management about their customers, about competing business, and about the environment in which they operate. The sensory capacity of the business to determine what is happening around it has improved and will continue to improve. Beyond this, the ability to distill nuggets of pertinent information out of such data enhances not only decision- making capacity but the goals on which such decisions are based. The secret here is not knowledge about customers, competitors, and the environment real knowledge about such matters is restricted to the

OCR for page 154
COMPUTERS AND BUSINESS 165 past but rather the quantification of uncertainty and risk. New analytic techniques will facilitate the introduction of uncertainty into decision processes, and from the beginnings that are already in place today, corporate actions will be weighed not only on the basis of expected return but on levels of acceptable risk. Here is an example. Suppose that a forecast of demand for an existing product has just been made. Using probabilistic tools, the forecast recognizes irresolvable uncertainties: the potential entry of competitors, the emergence of a new technology that could overtake the product, the potential for a fad that could spark unprecedented demand. Such factors produce two scenarios for the future. Suppose further that the first scenario requires building a new plant and the second does not. Has quantification of the level of uncertainty helped make the decision about the plant? Of course. One could reason as follows. Case 1: I believe the first scenario and build a plant, but the second scenario occurs. Case 2: I believe the second scenario and do not build the plant, but the first scenario occurs. Clearly one situation is better than the other, and even in this simple example of risk analysis, quantification of uncertainty helped resolve the issue. For risk analysis to become very accurate and helpful, models must improve and be trusted by managers, and data about environmental factors, competition, and customers must be collected regularly. All of these developments are happening and will accelerate. Consequences The use of corporate intelligence gathering, not as espionage but as a routine and accepted business function, will increase. Decision making will become explicitly risk-conscious, and decisions will be evaluated not in terms of return on investment (ROI) but in terms of ROI probability distributions. Intuition and the "gut call" will remain, of course, but in a probabilistic context some high-nsk opportunities will be seen as worth taking, while some lower-risk options will be judged not worth it. For better or for worse, probabilistic methods will make decision making more explicit and management more self- conscious and more accountable for its performance. CONCLUSION This has inevitably been a very rapid tour of some of the more important intersections in the function technology impact space, but several more cells deserve some mention.

OCR for page 154
166 THEODORE J. GORDON Computer-Aided Design and Computer-Aided Manufacture (CADlCAM). The design shop and the shop floor are being modified in function and form to capture the advantages of efficiency, quality, and accuracy afforded by the new technology. CAD/CAM went from a $1-million industry in 1973 to $1.2 billion in 1983. It is expected to grow by a factor of 10 between 1985 and 1995. Electronic Mail. Imagine an electronic typewriter that is likely to be on the market in 5 years or so. It has a matrix printer, memory for several lines, built-in word processing, built-in spelling correction, and several other easy-to-accomplish software features. This typewriter is sold not as an office word processor or computer peripheral, but simply as a consumer- oriented portable. Its price is less than $100, and it is a standard gift to high school graduates about to leave for college. It is not difficult to imagine the sale of several million of these devices, perhaps several scores of millions, within the next 10 years. With a modem chip these machines can be plugged into standard telephone jacks. Electronic mail, for better or worse, will have arrived overnight. Group Decision Making. Automated voting machines currently exist (e.g., the CONSENSOR) that permit participants in a meeting to provide judgments in response to questions posed by a monitor. In some of these machines, an individual's input can be weighted on the basis of expertise or knowledge. With improved expert systems and artificial intelligence, group interactions can be computer-augmented. Individual weights can be set by experience or testing, and the group itself can be integrated with expert systems and judgments drawn particularly against the profile of issues being addressed. Also, of course, on-line data can be called up if necessary to provide background for the group as a whole. These techniques tend to subdue the normal psychological problems that accompany group interactions and to promote smoother, more precise, and probably more accurate decision making. The consequence will be that teams, even teams composed of individuals located at remote places, become more important, at the expense of the individual. The way business does its work is being profoundly and permanently changed. On the factory floor it is being changed by numerically controlled machines, analog and digital sensing devices, automated testing gear, design systems, material-handling systems, inventory control systems, and automated stockrooms. In the office changes are being driven by electronic filing, automated scheduling of meetings, direct access to inflation, word processing, and, soon, idea pro- cessing. The computer affects almost every job and every worker. The nature of the jobs, where they are done, the way they are accomplished, the expectations of job supervision, and the skills required to perform the tasks are all changing. Additionally, the

OCR for page 154
COMPUTERS AND BUSINESS 167 information available for doing tasks and the precision and timeliness with which they can be done are changing. Granted business processes are changing, but is business itself changing? After all, business takes raw materials, adds value, and sells products. At this level, is anything likely to be different? Early critics of computers~omputers using cards that warned against folding, mutilating, or spindlingwere concerned about the way that computers would regiment and standardize us all, force us into providing rigid inputs that computers could understand. Now it is clear that computers provide the ability to manipulate and track information, and a variety of individual needs can be easily accommodated. Computers do not standardize; they promote diversity. We are coming closer to the time when the user can design the product and when, in sensitive and high- quality work environments, the worker can change his or her environ- ment and utilize an array of information that makes that person's contribution unique. Through the computer, business gains diversity. Business may also gain responsibilities. The prescription to take raw materials, add value, and sell products may be too simple for a future age. For all of its elegance, accomplishments, and promise, the computer can cause human obsolescence, displacement of workers, and unemployment. The role that business will have in accommodating these discontinuities is ideological as well as economic, and it is far from clear how that role will evolve. Somewhere in this chaotic, complex, and uncertain mix lies business of the future. NOTES 1. J. B. Angell, S. C. Terry, and P. W. Barth. 1983. Silicon micromechanical devices. Scientific American, April:44 55. 2. Committee on Science, Engineenog, and Public Policy; National Academy of Sciences; National Academy of Engineering, and Institute of Medicine. Cognitive science and artificial intelligence. In Research Briefings 1983. Washington, D.C.: National Academy Press, p. 25. 3. S. A. Levitan, and C. M. Johnson. 1982. Future of work: does it belong to us or to the robots. Monthly Labor Review, Vol. 105:1~14. 4. lI. N. Fullerton. 1980. The 1995 labor force: a first look. Monthly Labor Review, Vol. 103, No. 12:11-21. 5. R. M. Garsson. 1983. Fast growing ATMs are now as ubiquitous as Xerox machines Georgia finn aims new computer at bank calling of ricers. American Banker, December 21, 1983: 8ff.