Part 11
There is yet another way we can ban the computer from membership in our human society. While human beings occasionally think they are machines, and Dr. Bruno Bettelheim has documented a case history of “Joey” who was so convinced that he was a machine that he had to keep himself plugged in to stay alive, no machine has yet demonstrated that it is consciously aware of itself, as human beings are.
Machines are, hopefully, objective. Consciousness seems to be subjective in the extreme; indeed, some feel that it is a thing one of us cannot hope to convey as intelligence to another and thus has no scientific importance. It is also noted that the thinking and learning processes can be carried out with no need for consciousness of what we are doing. An example given is that of the cyclist who learns, without being “aware” of the fact, that to turn his machine left he must first make a slight swing to the right in order to keep from falling outward during his left turn. This observation in itself is not final proof of the pudding, of course, unless we are aiming only to make a mechanical bike-rider, but many of our other actions are carried out more or less mechanically without calling attention to themselves. Just as certainly, however, the thing called consciousness plays a vital role in human thinking. Perhaps the machine must learn to do this before it can be truly creative.
Although we have described some fairly “exotic” devices, it should be remembered that the computers in use outside of the laboratory today are fairly old-fashioned second-generation models. They have progressed from vacuum tubes or mechanical relays to “solid-state” components. When Artrons and neuristors and memistors and other more sophisticated parts are standard, we can look for a vast increase in the brain power of computers.
The Gilfillan radar ground-controlled-approach system for aircraft that “sees” the plane on the radar scope, computes the proper path for it to follow, and then selects the right voice commands from a stored-tape memory seems to be thinking and acting already. The addition of eyes and ears plus limbs and locomotion to the computer, foreseen now in the photocell eyes of Perceptron, the ears of Cybertron, and dexterity of Mobot and Hand, will move the computer from mere brain to robot.
Some people profess to worry about what will happen when the computer itself realizes that it is thinking, calling to mind the apocryphal story of the machine that was asked if there was a God. After brief cogitation, it said, “_Now_ there is.” To offset such a chilling possibility, it is comforting to recall the post-office electronic brain that mistook the Christmas seals on packages for foreign stamps, and the Army computer that ordered millions of dollars worth of supplies that weren’t needed. Or perhaps it isn’t comforting, at that!
The question of whether or not a computer actually thinks is still a controversy, though not as much so as it was a few years ago. The computer looks and acts as if it is thinking, but the true scientist prefers to reserve judgment in the spirit of one shown a black sheep some distance away. “_This_ side is black,” he admitted, “but let’s investigate further.”
------------------------------------------------------------------------
“_For forms of government let fools contest, That which is best administered is best._”
—Pope
7: Uncle Sam’s Computers
The modern electronic version of the computer is about fifteen years old, and like most teen-agers, it is a precocious child. To list all the applications in which it has made a place for itself would take several pages and an inclusive listing from Airlines to Zoology. There are hundreds of different types, priced from less than one hundred dollars to more than ten millions. The latter are so expensive that outright purchase is not usually possible for users. Rental or leasing arrangements are therefore available; and there are a growing number of computer centers to which the customer can take or send his work and have it done. There are also do-it-yourself computer facilities, much like those for laundry, dry cleaning, and so forth, as well as installations in trailers that move from place to place. Most require a source of conventional electric power, but there are some portable models that operate on batteries.
Scanning the list of jobs the computer now does, it would seem impossible to classify the varied tasks. Since many machines are versatile, general-purpose types, it is even more difficult to definitely categorize the computer. Dr. John R. Pierce, an expert at the Bell Telephone Laboratories, describes some of the chores done by a digital computer in a typical session at Bell:
Check parts of a computer program used in connection with machine methods for processing manufacturing information.
Process and analyze data on telephone transmission which have been transmitted to the laboratories by teletypewriter and automatically punched on cards for computer processing.
Solve a partial differential equation.
Compute details of the earth’s magnetic field.
Check part of a program used to handle programming cards.
Fit curves to data by translating numerical information into graphs.
Locate an error in a program designed to process psychological data.
These “simple” problems required but three minutes of the computer’s time. A larger task, something like solving 350 mathematical logic theorems from _Principia Mathematica_, takes a bit longer—eight and a half minutes, to be exact.
Despite this versatility, it is generally possible to break the computer’s capabilities down into broad classifications. First we can say that it does either simple data-processing, or scientific computations. Each of these can then be further subdivided ad infinitum. Examples will be seen as we describe uses of computers on the following pages. Since the government was the first user of computers, beginning back in 1890 with Hollerith’s punched-card machines, we would do well to see what other work it has put the computer to in the years that have elapsed.
_The Computer in Washington_
An inventory of electronic computers installed in the Federal Government by the end of 1961 totaled 800, with 200 more on order. These figures are exclusive of those for tactical and classified use by the Department of Defense. Some 45,000 people are engaged in computer operations in the government, and a total expenditure of about $1.5 billion is estimated. An indication of the importance accorded the computer by Washington is the Interagency Data Processing Committee, concerned with questions of sharing of computers in geographic areas, setting up of a “library” of applications, and assurance of continued computer operation in the event of attack or other emergency. Users of computers, in addition to the Department of Defense, are the Atomic Energy Commission, Department of Commerce, National Aeronautics and Space Agency, Federal Aviation Agency, Post Office Department, and others for a total of 43 agencies. The Peace Corps, for example, recently announced that it would acquire a computer for use in its work.
Red tape is not the only output from Washington, D.C. Not long ago the Hoover Commission estimated that our Federal Government also produces 25 _billion_ pieces of paper each year! Someone else converted this already impressive statistic into the more startling information that placed end to end these papers would reach the moon four times—in triplicate, of course! Data-processing, then, the handling of information, would seem to be the major part of the computer’s work for Uncle Sam.
The Census Bureau was the first government user of the computer, and it continues to handle its work in this way. In 1951 the government procured a UNIVAC I to take over this onerous chore from its predecessors. Beginning with the 1950 census, the computer has been in operation practically twenty-four hours a day, seven days a week. In its first ten years it performed more than 510 billion mathematical operations in keeping pace with our exploding population. We are producing more than paperwork, it seems. The 1950 census required four years to process. With newer computers the 1960 count will take only half as long despite the population explosion.
Information-handling computers make possible another important phase of the government’s work. In 1936, machines began to process Social Security records, which are becoming a monumental pile of paperwork themselves with close to 100 million accounts that must be kept up to date. Social Security numbers recently turned up in government computers handling another job—that of income-tax bookkeeping.
The U.S. Commissioner of Internal Revenue, Mortimer Caplin, put a pilot system of computer accounting of tax records into operation in January of 1962 in the Atlanta region. In 1963, the Philadelphia Center will follow suit, and by 1966 all income-tax accounting will whiz through tape reels into computers. The figures on tax greenbacks laid end to end are not available, but it is known that 400 miles of magnetic tape will be needed to hold all the records.
The new system will make it tough on the income-tax chiseler. Caplin points out that not only the withholding-tax information from the employer, and forms from the employee, but also dividend statements and other supplementary income information will funnel automatically into John Doe’s portion of the tape. If John is moonlighting, holding down a second job he might forget to mention, the computer will spot it and charge a tax on it. The apprehended tax-dodger may well call the computer an infernal revenue machine.
There are of course many other ways the computer is helping out in the complex problems of government, both Federal and local. The computer has already figured in national elections, making predictions well in advance as to the outcome. Now the machines are being used in the actual voting procedure. In 1952 an IBM computer predicted Eisenhower’s victory within two hours after the first polls closed. In the early days of computer predictions, the men using them were overly cautious and afraid to accept the machine’s word. Techniques and confidence have improved with practice, and in 1960 IBM’s RAMAC predicted victory for Kennedy at 8:12 P.M. election night.
To make accurate predictions, the computer is given information from preceding elections. In 1960 it was fed the results of the 1956, 1952, 1948, and 1928 (because of the religious considerations) elections. Forecasters were able to ask the computer such questions as, “How is Nixon doing compared with Eisenhower’s showing in 1952?” “How is Kennedy doing compared with Al Smith back in 1928?” “Is labor voting as a bloc?” and “How solid is the South?” The computer is now an accepted part of network equipment for election reporting. ABC used the Remington Rand UNIVAC; CBS, IBM RAMAC and other machines; and NBC the RCA 501.
[Illustration:
_International Business Machines Corp._
Computers are used to predict the results of elections. ]
In addition to forecasting results, computers are beginning to do other election work. Los Angeles County experimented with a computer method of counting votes in 1960. Greene County, Ohio, used punched cards for ballots for 50,000 voters in a pioneering computer voting system. The cards were processed in a UNIVAC computer at Dayton Air Force Depot. A bolder suggestion is that of political scientist R. M. Goldman of Michigan State University: actual voting by telephone-operated computer!
To solve another problem area in voting, the use of computers was recently proposed at a state congressional hearing in Boston. Redistricting, the bugaboo that led to “gerrymandering,” might well be done by “unbiased” computers which would arrive at an optimum redistricting plan. These unbiased results would be “beyond politics and in the best interests of the voters and the State,” according to the computer expert who proposed the plan.
Moving from voting to a more complicated problem, that of urban renewal, the University of Washington is conducting a survey under federal grant on the extent of deterioration and the causes of decay in Spokane residential, commercial, and industrial areas. The IBM 709 computer makes possible an accurate and extensive survey expected to shed light on areas of arrested development, and on the amount of tax revenue lost because of existing blight.
_Electronic Legal Eagle_
Some writers see the clearest evidence of the victory of the computer—if indeed we admit to there ever having been any real battle—in the admission by the legal profession that it must begin to chart the legal seas of the computer age.
In 1961 the American Law Institute and the American Bar Association, feeling that the computer will cast its “automated shadow on every phase of society,” conducted a joint three-day seminar in Washington, D.C. Titled “Legal Problems in the Use of Electronic Data Processing in Business, Industry and Law,” the seminar discussed “function and operation of computers and their impact on tort, tax, corporation, labor, contract, banking, sales, antitrust, patent and copyright law, as well as on the law of evidence and trial practice.”
Lawyer Roy Freed of the Philadelphia Bar, in a booklet called “A Lawyer’s Guide Through the Computer Maze,” describes the working of the machine and then poses some challenging legal questions.
What duty does the company acting as a computer service organization have to preserve the confidential nature of the data it processes for its customers?
Can business records placed on magnetic tape be used in evidence, or must the original records be preserved?
How long can corporate management lag behind others in their industry in adopting machine data-processing systems before they expose themselves to a mismanagement charge?
To what extent should the manufacturer of a complex product that has a potential for causing harm try to minimize his liability as a maker by anticipating design defects through simulated operation on the computer?
Other legal experts asked other questions. If an electronically processed check is bounced erroneously, who is responsible? If a noncomputerized railroad has a train wreck, can the road be sued on the premise that the accident would not have occurred with modern traffic controls? Or if the reverse occurs, can an anticomputer claimant win a suit against the machine?
Applications of the computer in patent law may lead to more thorough search in addition to higher speed. This could well clear another bottleneck by issuing fewer and faster patents. But copyright violation problems lie in the possibility of making copies of tapes or other media suitable for the computer’s use. The altering or falsification of computer data also poses a tricky legal problem; there is already a precedent in the Wall Street man who juggled the punched cards on the computer to his own advantage.
Perhaps there was one question none of the lawyers present had the courage to bring up: what if the day comes when the court itself is a computer, and the case is presented to it as a stack of cards, or a prepunched or magnetized tape? Such a mechanized justice was fancifully depicted on a television thriller by Ray Bradbury.
_Computers in Khaki_
Despite the low IQ it has been accused of, it was inevitable that the computer be drafted. In the 40’s we were desperate. Included in government use of computers are military research, development, and tactical and strategic methods. World War II was a different kind of war, a complicated, electronic war that required advanced methods of operation. At Eastertime in 1942, IBM answered an urgent call from Washington and gathered all available data-processing machines for use by the military. Punched cards kept track of allotments, insurance, and the logistics of running a war. Mobile computing machines operated close to the front lines, and were important enough that a captured German officer was carrying urgent orders to bring in one of these units.
[Illustration:
_Motorola, Inc., Military Electronics Division_
Technician checks circuitry of airborne digital computer. ]
Besides the mundane effort of mere data-processing, wartime computers did important cloak-and-dagger work as well. A report came in from Allied intelligence that the Germans were working on a frightening new development—an electrically powered cannon. If it were successful we would need some kind of counterweapon. But the dike was leaking in a hundred other places too, and there was not time or equipment to do everything it seemed we might have to do. The answer was to feed some complex mathematics to an IBM computer called the Automatic Sequence Controlled Calculator—mathematics describing the new cannon. The computer cogitated briefly and decided that the Germans were on the wrong track; that the gun would not work. We therefore ignored the threat, letting the Germans waste their valuable time going down the blind alley, and turned our efforts elsewhere.
We have said that World War II was a different kind of war. One new development to bear out this difference was called “Operations Research”—the reduction of any program to mathematical formulas and the investigation of these formulas rather than a conventional, intuitive approach. The technique was pioneered in England, spread to the United States, and is now one of the most powerful tools not only of the military but also of government and business. The computer has made operations research a more powerful technique by permitting the analysis of thousands or millions of possibilities in hours instead of lifetimes.
Back in the days of bows and arrows, the soldier had no need for a computer. Even the rifleman required little more than a simple sight and maybe a bit of Kentucky windage. With the coming of long-range artillery, computers became desirable, and now we have moved into an age of warfare that would be impossible without high-speed computing machines.
In 1948 IBM introduced a computer known as SSEC for Selective Sequence Electronic Calculator. This machine was put to work on a problem for the Los Alamos Atomic Energy Laboratory, a problem called “Hippo.” Hippo was as unwieldy as its name, requiring some nine million involved mathematical operations that would have taken about 1,500 man-years of skilled time. That many mathematicians or that much time was not available, of course, and SSEC clicked through the job in 150 hours by itself. Another computer, the MANIAC, designed by John von Neumann, is credited with beating the Russians to the punch with the hydrogen bomb.
As an outgrowth of operations research, the simulation of war games has become an important part of military work. A number of firms, including System Development Corporation, Technical Operations, Inc., and others, devote much of their time to “playing games” to work out the optimum strategy and tactics for war in case we find it necessary again.
It is perhaps not paradoxical that war be considered a game. As William Cowper said, “War’s a game, which were their subjects wise, Kings would not play at.” The game of chess, conversely, stems from war and its tactics. Indeed, the term checkmate, for victory, comes from the Persian words _shah mat_, the king is dead.
Through the years many war games have been developed, games which eliminate the physical conflict but preserve the intellectual maneuvering necessary for waging “war.” John von Neumann was one of the more recent to turn his great genius to this subject in the development of his “minimax” theory. This is an outline of a situation in which consequences of decisions depend on the actions of an opponent. We have seen that the computer, though not yet world champion, can play chess; the minimax theory is more grist for its electronic mill.
Tech-ops operates the Combat Operations Research Group for the U.S. Continental Army Command at Fort Monroe, Virginia. Among the games played here with computers are SYNTAC, in which field-experienced officers evaluate new weapons and tactics, and AUTOTAG, a computer simulation of tank-antitank combat. Other projects of this firm include air battle simulations, ship loading and other logistics problems, fallout studies, and defense against missile attacks. The beauty of such schemes is that we will not make the mistake of the Germans with their electric cannon. When the computer blinks “Tilt” or an equivalent, the engineers may have red faces, but no huge amount of time or money will have been spent before they sigh, “Back to the old drawing board!”
[Illustration:
_Aeronutronic Division, Ford Motor Co._
ARTOC (Army Tactical Operations Central) uses computer techniques for battlefield display and communications to aid field commanders. ]
At Picatinny Arsenal, computers evaluate ammunition by simulating as many as a thousand battles per item. Design and management studies for projects like Nike-Zeus and Davy Crockett are also conducted at Picatinny. A mobile computer, called MOBIDIC, is designed for field combat use and has been moved in three 30-foot trailers to location with the Seventh Army in Europe. There it handles requisitions for rockets, guided missiles, electronic equipment, and other items. MOBIDIC is just part of the Army’s FIELDATA family of computers that includes helicopter-transported equipment to provide field commanders with fast and accurate data on which to base their risk decisions. Another concept is ARTOC, for Army Tactical Operations Central, an inflatable command post in which computers receive and process information for display on large screens. This is a project of Aeronutronic.
In 1961 an IBM 7090 computer was installed at Ispra, Italy, for use by the European Atomic Energy Commission (EURATOM). The computer would have as its duties the cataloging of technical information on atomic energy, the translation of technical publications, and use in basic research on solutions of Boltzmann equations and other advanced physics used in atomic work. In this country, the National Science Foundation has acknowledged the importance of the computer in scientific investigation by underwriting costs for such equipment for research centers in need of them.
[Illustration:
_International Business Machines Corp._
Command post of SAGE, the most complex computer application to date. ]
_In the Air_
Beyond the realm of war gaming, the computer also forms the heart of the hardware that such simulation and studies develop. SAGE is an example of this, a complex warning system that protects our country from attack. The acronym SAGE is a more dignified and impressive name than the words it stands for—Semi-Automatic Ground Environment, an environment that by 1965 will have cost $61 billion!
Sage is not a single installation, but a vast complex of centers feeding information from Ballistic Missile Early Warning Site radar and airborne radar, from ships, Texas towers, and ground-based radar, and from weather stations into a central control. This control sends the proper signals to defensive rockets, missiles, and aircraft for action against an invader. It does this with one hand, while with the other it keeps tabs on normal military and commercial air traffic.
The System Development Corporation designed and IBM built the SAGE computer, a computer already old-fashioned since it uses vacuum tubes instead of the newer transistor devices. Despite this shortcoming, it does a fantastic job of tracking all the aircraft and missiles in its ken, labeling them for speed, heading, altitude, as well as the vital information of friend or foe, and continuously plans a defense. Since it can monitor civilian traffic as well, SAGE may one day take over control of that too. Thus the money spent will yield a bonus in addition to the protection SAGE has already afforded in its military role.
The Air Force uses airborne computers by the thousands. Indeed, the need for small lightweight computers for applications in aircraft led to early work in the miniaturization of components that made possible tiny computers for missile and space use. Small digital computers were built for “drone” aircraft navigation; now more advanced computers provide “air data,” air-speed, altitude, flight attitude, pressure, and other information.
Other Air Force computers, used in BMEWS radar, take the place of human observers. These smart computers can recognize radar tracks that are potential missile trajectories, discriminate among these tracks to select hostile trajectories, and project them to impact points and times. Called MIPS, for Missile Impact Predictor Set, the computer takes over from its human forerunner who just can’t seem to perform the 200,000 operations a second required to do the job.