The proper handling of requests, regardless of the identity of the requestor, deeply concerns many Federal agency officials. The Ruggles report states that: "Outside requests for data are often uninformed, unreasonable, and in view of the Federal agency, not worth while. Few outsiders can know enough about the data, their nature and characteristics to make sensible requests, or to have a realistic appreciation of the analytic limitations which the data impose." In summary, it would appear that a better understanding of what is involved in protecting and disclosing information is needed. Dunn emphasizes that "no number conveys any information by itself. It acquires meaning and significance only when compared with other numbers." 10 He then amplifies upon the crucial consideration concerning information disclosure by declaring: "The strain upon disclosure arises because matching several sets of data for consistency at levels of aggregation appropriate to the problem requires a retreat to elemental units in the process of constructing the necessary aggregates.” 99 11 Admittedly, the creation and enforcement of adequate safeguards may inhibit the random type of research "browsing" that is always in progress, but the expressed desire for information protection will necessitate certain compromises. SUMMARY Although the feasibility of establishing a Federal Data Center to serve as a repository for a wide range of economic data has been established, the desirability of this action still is in doubt. The value of centralized statistical data for use by government and private research and planning personnel vis-à-vis the alleged threat to citizen's privacy continues to be the subject of debate both within the Congress and in the public press. The state-of-the-art of automatic data processing today reflects spectacular advances in equipment and technique development. The issue now under consideration centers about the controls which must be delineated, evaluated, and implemented if such advanced devices and procedures are to be employed to the benefit of the greatest number of persons in these United States. THE PROPOSED FEDERAL DATA CENTER: SELECTED REFERENCES 1. Committee on the Preservation and Use of Economic Data (Chairman: Richard Ruggles), Report of the Committee on the Preservation and Use of Economic Data to the Social Science Research Council. Washington, D.C., April 1965. 43 p. plus Appendices. 2. Dunn, Edgar S., Jr. Review of proposal for a National Data Center. Office of Statistical Standards, Bureau of the Budget, Washington, D.C., December 1965. 40 p. plus Appendices. 3. Experts say computers could aid a "big brother". The New York Times, October 4, 1965. p. 120. 4. Horton, Frank. Congressman Horton questions proposed computer central for federal data. Congressional Record, June 14, 1966. pp. 12487-12488. 5. Horton cites potential privacy invasion by proposed central data system. Congressional Record, August 5, 1966. pp. A4143-4144. 6. Lardner, George, Jr. Center for data on everybody recommended. The Washington Post, June 13, 1966. p. A-3. 7. - Data center hearing warned on privacy. The Washington Post, July 27, 1966. pp. A-1 and A-8. 8. Long, Edward V. The dossier. Congressional Record, August 5, 1966. pp. 17560-17561. 9. 9603-9605. The ultimate big brother. Congressional Record, May 9, 1966. pp. 10. Pooling of U.S. data on individuals urged. The Evening Star, June 13, 1966. pp. A-1 and A-6. 11. Too personal by far. 12. To preserve privacy. The Wall Street Journal, August 5, 1966, p. 6. 9 Ruggles Report, op. cit., p. 26. 77-577-67-11 COMMUNICATIONS, COMPUTERS AND PEOPLE (By Paul Baran, the Rand Corp., Santa Monica, Calif.)1 INTRODUCTION My friends in the computer business sometimes refer to me as a "communicator" and those in communications speak of me as being a "computer-type." I don't think that these words are meant to be derisive, but it does put me somewhat outside or at least between the two separate and individually respectable fields of computers and communications. But I expect time will bail me out by blending together the two separate technologies into one. Merely by standing still, I will be able to welcome both sets of friends as they join me in this small but rapidly growing joint field being created by a dual technological development. Communications and computers are today becoming what the economists call "complementary goods"-one without the other is of much lesser value-like pen and ink, pretzels and beer, and gin and dry vermouth. ORDER OF PRESENTATION We shall first briefly discuss the impact of the computer technology upon the communications business and, conversely, how good, widespread, low-cost digital communications will allow a dramatic increase in the creation of new types of computer systems. Then we shall get down to the meat of the talk-a few of the unappreciated social consequences possible and, lastly, to proffer remedies in advance of the time society realizes there is a problem. If the order of things appears backward, with remedies being offered in advance of the patient's complaining of an ailment, we do so with the belief that the lead time for the cure of social ills is often longer than the gestation period of the disease. We alone who appreciate what is happening to computer development may be in the best position to see the thunder clouds. THE IMPACT OF COMPUTER TECHNOLOGY ON COMMUNICATIONS Communications equipment is sometimes categorized into switching equipment, transmission equipment, or terminals. As we expect to be talking about digital uses which by definition are digital terminals, we shall confine our observation to telephone switching and transmission. At present, the telephone plant is almost exclusively based upon electromechanical switching-that most primitive form of computer logic-and one that we in the computer business haven't seen around for years. Transmission is by means of frequency division multiplexing of analog voice channels requiring oscillators and modulators terminating into banks and banks of hetrodyne circuits and sharply tuned filters or about as analog (or undigital) an operation as we computer types can envision. The only kind words a computer man can have for this system is that it works; it works well and has been working well for many years. While perhaps slow by pace, electronic switching has arrived on the scene for the telephone company. At least two separate systems in this country have now passed field trials and are being installed commercially. This new equipment may be representative of the future telephone local central offices. At present, these electronic switches are not believed to be more economical than their earler electro-mechanical switch counterparts. But their prime advantage lies in the new additional services that they offer because of the general computer nature of the control mechanism of the switching center. For example, it will be possible to dial only two digits to reach the few numbers that you call often. It will be possible to relay a call to another telephone if you are temporarily away. Automatic diagnostic routines will permit repair and maintenance by inexperienced personnel. Further, electronic switching is a new technology whose price is expected to decline rapidly in the future. So much for switching. We also see computer technology creeping into the picture in transmission. The Bell System T-1 multiplexing system samples 24 1 Any views expressed in this paper are those of the author. They should not be interpreted as reflecting the views of The RAND Corporation or the official opinion or policy of any of its governmental or private research sponsors. Papers are reproduced by The RAND Corporation as a courtesy to members of its staff. This is an invited paper scheduled for presentation at the Fall Joint Computer Conference to be held in Las Vegas, December 2, 1965. analog voice channels about 8,000 times per second producing, together with synchronizing information, a data stream of 1.54 megabits per second which can be transmitted over ordinary copper pairs using pulse regenerative amplifiers. This pulse code modulation technique is being developed simultaneously in many countries and is in use but presently restricted for links on the order of magnitude of about 20 miles. As pulse code modulation is the most economical of the multiplexing systems, it appears destined as the transmission direction of the future. Even though digital technology is entering the telephone plant slowly and in a piecemeal fashion, it is arriving and will make its impact felt. Specifically, most of the growth of the telephone plant may be expected to occur within these digital techniques areas. The implications to us are several-fold. First, as we expect to see a rapid drop in the cost of digital circuits, we may expect continued drops in the price of digital communications in the future. We would also expect to see even more marked savings to the digital communicator as systems evolve which are more amenable to the all-digital processing of information from user to user. Today emphasis must be given to complete compatability with the large existing analog system in being where periodic reconversion to analog signals is required. Thus, one day we envision the bulk of the telephone system being built entirely of digital processing assemblies in lieu of the all-analog systems as of today. When this day comes, we computer types would view the telephone system as merely being another particular computer application and not necessarily a specialty field unto itself. If the old-time telephone engineer fears that the computer types are taking over, he is probably right. So much for what computer technology might do for or to communications, depending where you sit. THE IMPACT OF COMMUNICATIONS UPON COMPUTERS Using telephone lines modified to handle digital data, we are able to build an increasing number of geographically distributed time-shared computer systems. Many individual users are connected to a common computer data base. Examples of such systems include airline reservation systems for civilians and fancy display "command and control" systems for the military. Simple record keeping, a mark of a highly developed economy, has been a prime area of development of these large computer file/communications systems where much of the routine clerical work is transferred to the computer with human interrogation of the system. As time moves on, the number of people who will be able to interrogate the system and the geographical distance between them and the machine will increase. SOME INDIVIDUALLY USEFUL SYSTEMS Today we see time-shared file systems used for insurance records, for checking automobile tags, to locate outstanding criminal warrants, and for credit check investigations (using drivers' license numbers) in cashing checks. The systems built to date pose no overt social problem. The information handled is not highly sensitive and access to it is generally limited. THE TRAIL OF ARTIFACTS IN A CIVILIZED LIFE As we pass through life, we leave a trail of records widely dispersed and generally inaccessible except with a great deal of effort and diligence. We start with a duly recorded birth certificate. We leave behind hospital records and our pediatrician adds to our medical records. We are deductions on our parents' income tax. School is a place where we busily generate record upon record of our scholastic grades, our attendance, our IQ test records, our personality profile records, volumes galore. With automated teaching coming to the fore, we can expect better record keeping. The volume of data we will record per child may be expected to increase even more markedly ("in the best interest of the student"). Between terms we get our social security card and a job, and we start leaving behind us a long history of employment records. We reach age eighteen and are entered upon the records of the Selective Service. We get a driver's license and, if we are lucky, we will be able to avoid having arrest and jail records. Most of us will apply for a marriage license, some of us will collect divorce decrees which will end in voluminous court records. We move from job to job in a mobile economy creating moving-company inventory records of our goods. Even as we move from place to place we leave behind short records of our airplane reservations and for some reason every hotel makes a ritual of acquiring and preserving the names and addresses of its guests for posterity. This list is only a partial one. Play the game yourself and think of all the records you leave as you go through life. WHY SO MANY RECORDS? One does not create records merely for the sake of creating records. But rather there is the implicit assumption that the records will be of some use some day. In order to be of use, there must be some means of interrogating the files to resurrect the information sought. Thus, we envision large families of systems, each individually useful. For example, an Internal Revenue Department investigator might wish to have immediate access to the tax returns of the associates of a man who is being audited to check for consistency of financial relationships. A company may wish to have rapid access to its personnel files to know whether to give a good reference to a former employee. A doctor may wish to trace the entire medical history of a patient to provide better input into a diagnostic computer. The Veterans' Administration may wish to examine a man's complete military record and possible other previous medical records to see whether the ailment claimed as being service-connected really is. A lawyer for the defense of a man will wish to search for jail records, arrest records, and possibly credit records of all witnesses for the plaintiff. Professional licensing boards may wish to delve into any records that may indicate the applicant lacks an unblemished character. The military in filling extremely sensitive positions may even wish a record of all books borrowed by the prospective applicant to insure that his interests are wholesome and he possesses the proper political bias desired. ACCESS TO THIS INFORMATION Today one does not gather such information about the prospective examinee easily. If one went through the direct channels and asked most sources for their records about a person, he would most likely be told to go jump in a lake, if for no other reason than the information is not available-cheaply. Even if the infor mation were a publicly available record, the investigator must be expected to spend a great deal of time and effort delving through to discover pertinent data. Today, as a practical matter, if one wishes to obtain much of this information about a person, he hires a private detective who charges a great deal of money and expends a great amount of time obtaining a little information available from a portion of these potential records. The price for a fishing expedition for infor mation is high and most of the fish are inaccessible. THE IMPENDING PROBLEM So much for the pleasant past. Consider the following argument: 1. A multiplicity of large remote-access computer systems, if interconnected, can pose the danger of loss of the individual's right to privacy-as we know it today. 2. The composite information data base may be so large and so easily acces sible that it would permit unscrupulous individuals to use this information for unlawful means. 3. Modern organized crime should be expected to have the financial resources and access to the skills necessary to acquire and misuse the information in some of the systems now being considered. 4. We are concerned not only with the creation of simple "automated blackmail machines” using this information, but with the added implication of the new "inferential relational retrieval" techniques now being developed. Such techniques, when full refined, could draw chains of relationships from any person, organization, event, etc., to any other person, organization, or event. 5. Humans, by their day-to-day necessity of making decisions using totally inadequate evidence, are innately prone to jump to conclusions when presented with very thin chains of inferred relationships. For example, merely plastering a man's name on billboards will markedly change the outcome of an election, if the other candidate's name is not equally displayed. 6. The use of private detectives to unearth derogatory information on political candidates and their associates has become an increasingly prevalent feature of elections. This practice is expected to increase in the future. 7. The cost-per-unit-dirt mined by unautomated human garbage collectors can be cut by orders of magnitude once they obtain access to a set of wide-access information systems which we now see being developed. It is the sophisticated form of chain-relation blackmail that may be of most social concern. We generally pass through three stages of information storage development. First, we start by keeping manual records employing clerks. Next, we get rid of some of the clerks when we put all the records into a single central computer file with the readout controlled from a single point. The next step is the creation of remote interrogation devices to interact with the file from a large number of points. The payoff for instant access is often high as it eliminates all delay to the file user. 8. This development of geographically widespread access systems requires the use of communications lines to connect the users into the computer. There is a widespread belief that somehow the communications network used will possess a God-given sanctuary to privacy, but "it ain't necessarily so. DIRECTIONS TOWARD A SOLUTION 1. Assume that not everyone is as honest and as trustworthy as ourselves— but is just as diabolically clever. 2. Appreciate that we will be increasingly dealing with complex and, hence, difficult-to-understand-all-the-details types of systems in the future. 3. Probably the only people who best understand the operation of each system will be computer design engineers who build the system in the first place. 4. The only time that the fundamental safeguards that we seek can often be applied is at the time of the initial system design. "Software patch-ups" at a later date may generally be relatively ineffectual compared to good initial design-good design being defined as including an awareness of the existence and importance of the problem. 5. Do not expect help from the legal profession in lieu of good design. Even ignoring the social lag of the legislative/judicial procedure, the detailed subject matter verges on or beyond the limits of their comprehension. 6. Laws and laws alone have been almost totally ineffectual in the growth of widespread electronic eavesdropping and wiretapping. At most, all the courts have accomplished is to prevent the police from using the same techniques available to the private detective or the criminal-or even casual readers of an electronics technician magazine. 7. While I have little faith that laws in themselves will solve the problem, laws could be helpful in two ways: a) Laws outlawing certain practices will be of minor help in increasing the price of the act and making its commission less flagrant; b) laws can be written so that potentially weak systems cannot be built unless adequate safeguards are incorporated throughout for the protection of the information stored. 8. This last direction is to me viscerally unsatisfying as it carries with it a built-in loss of freedom. The thought of the creation of another governmental agency peering over one's shoulder contains the seeds of the possibility of bureaucratic decay and arbitrary conclusions based upon an incomplete understanding of complex problems. 9. Historically, government regulatory agencies start as highly effective bodies but lose momentum as the original personnel leave and their replacements come from the industry being regulated. (Where else are you going to get competent people who know the business?) The extreme competence that we need in a regulatory agency of this type is too rare a commodity. 10. If we are to avoid external regulation, then it behooves us computercommunication system designers to start working, or at least thinking about the problem. We should take the initiative and the responsibility of building-in the needed safeguard ourselves before Big Brother is forced to do it himself and we are not too happy with the way he might want to do it. 11. Safeguards, whether they be screens around moving machinery or circuit breakers, cost money. Every design engineer is reluctant to add anything that costs money and buys little visible protection. But the writer believes that it is time to start regarding such added costs as necessary costs—a price to society for the privilege of building a potentially dangerous system. 12. This is not a new concept. We have, for example, been practicing this in the design of sewerage systems and in electrical distribution systems for some time. But, historically, it has generally taken an epidemic to build a local sewer |