CANADIAN BAR ASSOCIATION ANNUAL MEETING.
The principle of using computers to assist the legal profession is no longer questioned. In fact, there are so many areas in which computers are used in the aid of justice that it is easy to overlook some of the more important and more imaginative areas.
Much of the work to date has involved the use of computers in legal information retrieval. This area is reasonably well advanced and will be the subject of the talks of other speakers.
I felt it would be more useful to review some of the frontier areas with a view to stimulating further investigation of the use of computers.
I don’t pretend to be an expert in these areas, nor do I claim that all of the uses referenced will in the long-run prove to be equally fruitful. This is for you to decide.
Justice Douglas said, in 1948, “The law is not a series of calculating machines where answers come tumbling out when the right levers are pushed”. However, in 1964, the Justice recanted somewhat, and said, “that he does not mind if computers are used to help make decisions, just so long as he gets to push the buttons”.
Justice Douglas had correctly recognized that the success of the use of computers to assist the legal profession will depend on the involvement of lawyers in this process. With that in mind, let us now survey the field.
1. Collecting the Evidence
It is often forgotten that computers start to assist the cause of justice long before a case reaches court. Computers are used in some unusual ways for crime prevention.
For example, considerable work has been done on the division of police districts into patrol beats. Standard techniques of operations research are used to analyze the number of police required per 1,000 population (4), how the manpower should be allocated according to level of crime by type of district, what areas are best patrolled by car, on foot, or on scooter patrol, as well as the long-range manpower planning. One of the problems this type of analysis introduces is the necessity of being able to refer to small areas of the city. This normally involves a form of geo-coding such as being undertaken by the Dominion Bureau of Statistics and others, to assist in census taking in large cities. (Saul I. Gass, IBM Federal Systems Division, ACM Proceedings, 23rd National Conference, page 459.)
This application is only one example of the use of the technique of computer simulation in a narrow aspect of police work. The problem facing North America was graphically pointed out in Edgar J. Hoover’s recently released Statistics on Crime in the United States. Last year, the chance of an individual being the object of some attack on the person, e.g. robbery, assault, rape, etc. was one in fifty. Today’s police force cannot come close to coping with this situation without modern techniques to help optimize their war on crime.
Another example of a common computer application to assist the police is in the maintaining and searching of a modus operandi file. The M.O. file is built up by recording on a standard code sheet, the particulars of the crime as to:
â€¢ time of attack, e.g. funeral, weekend, holiday
â€¢ object of attack, e.g. auto accessories, cameras, cash registers, etc.
â€¢ type of property attacked, e.g. airplane, auto, boat,, church, school
â€¢ how attacked, e.g. breaks glass, cuts glass, etc.
â€¢ trademark, e.g. ate on premises, barefoot, etc.
â€¢ vehicle used, e.g. stolen auto, rented car, etc.
These facts, stored in a computer, are then searched to establish a list of suspects for an unsolved crime, to determine that a series of crimes may have been committed by an individual arrested for one particular crime or to determine that a series of unsolved crimes appear to have been committed by the same unknown criminal thus guiding future investigation of the crimes. (Thomas H. Giske, IBM Sacramento, Computers and Automation, February 1963)
It is interesting to note that extensive work in this area is being done in other countries including the Soviet Union. An article in Cybernetics and Law, Moscow 1967, the Russians report considerable work in the study of handwriting for the apprehension of criminals. This is used to buttress expert opinion. Other uses under study by the Russians include analysis pointing to similarity in prose.
â€¢ The Russians, like ourselves, are doing considerable work in the area of fingerprint analysis on computers, as well as the comparative analysis of photographs of faces. One might suspect here that such investigations have more to do with an individual dossier intelligent system than crime control in our sense of the word.
The reason for mentioning this area is that it does point up another frontier computer application in criminology and that is ‘pattern recognitionâ€™.
We should reference in passing, that computers are used extensively in other police administrative matters, such as motor vehicle licence registration. The mere availability of such masses of information on computers, does assist greatly in the prediction and apprehension of criminals. I have no doubt that one of the other speakers will reference the problems of privacy when databanks of such information are established about individuals.
One could go on to speculate that even such things as traffic control systems, which are now computerized in Toronto and else where, could be used to assist in apprehending criminals, although I know of no instance of this being done to date. Computers are certainly used in this area for accident prevention.
In general, then, the major computer techniques used to assist police are simulation, pattern recognition and information retrieval.
2. Analyzing the Evidence
Having decided to bring the action to court, there is the problem of putting together and assessing the mass of evidence necessary to make a good case. Once again, computers can be of assistance.
Some civil cases involve masses of data which must be correlated and be available for retrieval before and during the court presentation. A recent example was the Peace River Dam lawsuit in British Columbia. Mr. Justice J.A. Macdonald of the Supreme Court said that thousands of documents will be involved, including 3,600 construction drawings, 60 volumes of letters, and large numbers of memos and photographs”. Douglas Brown, counsel for prosecution against BC Hydro, said that a computer will be used to sort out the documents – the first such use in Canadian legal history, to his knowledge. (Vancouver Province – June 12, 1969)
Once the evidence is assembled, however, it should be possible to use computers to assist in assessing the logic and consistency of evidence. In fact, this field of jurimetrics, because of its use of symbolic logic, lends itself very well to the use of computers. P. Meyer, in the Canadian Bar Review, March 1966, reviewed some of the techniques. In general, it should be possible, in complex cases, to use a computer to sort and sift information to determine the correctness of time sequences, the consistency of alibis, and other such uses, particularly where court cases drag on over many months with massive amounts of testimony.
â€¢ The use of computers in this area involves the more standard techniques of data retrieval and data manipulation.
3. Predicting What Will Happen
Leaving aside, for the moment, the obvious use of computers to search for legal precedents, there has been some preliminary work done on the use of computers to predict judges’ decisions. Reed C. Lawler, a Los Angeles Patent attorney, has been working for almost ten years in an effort to describe some phases of the judicial decision making process by means of mathematical formula. Much of this work was done studying appellate court decisions in a project with the University of Southern California Law Centre.
Lawler hastens to point out that computers will never replace judges in handing down decisions. The work of people in this field has given rise to speculation that if you could always predict a judge’s decision, eventually you do not need the judge.
Miss Sally Dennis of IBM, pointed out in the January 1968 issue of Law and Computer Technology that judges should be complimented rather than alarmed. At least this indicates that judges’ decisions are sufficiently consistent that some prediction should be possible.
A judge’s attitudes tend to program him in somewhat the way one programs a computer. Lawler’s approach is to reduce the pattern of facts to mathematical formula and then match these against the pattern of facts of previous cases decided by that judge. He proceeds on the assumption that every fact is favourable to either a positive or negative answer to the issue posed in the case. The fact pattern is then reduced to a string of binary ones and zeros.
I make no attempt to assess what Lawler calls ‘Personal Stare Decisis’ where simple precedent is replaced by the judge’s decision making patterns.
It is interesting to speculate, however, that the use of such techniques in advance of a case reaching a particular judge might cause a lawyer to advise his client whether to proceed with the litigation. (Law and Computer Technology, May 1968, page 11)
â€¢ Here the use of the computer’s logical ability is the major technique.
4. Creating the Law
Computers can be used in a variety of ways to assist legislators in creating the law. Some of these techniques are very straight forward.
For example, a modern legislature would be hard-put to handle the mass of law that must be handled in a single session without the assistance of a computer. (Although most Canadian legislative bodies are endeavouring to handle the workload manually.) For example, there were 26,566 bills and resolutions introduced before the 89th Congress in the United States. These bills went through innumerable references to committees, readings on the floor of either House, changes and alterations, etc. Of this number of bills, 4,016 were passed into law. (This can be contrasted to 118 bills passed by the first Congress of the United States.)
It is proposed that for the U.S. Congress, computers will be used to give:
â€¢ status of pending legislation
â€¢ current schedule of committee meetings and hearings
â€¢ even an up-to-date telephone listing
An analysis of the needs of Congressmen showed that 78% considered their major problem to be “complexity of decision making and lack of information”.
Vice President Humphrey commented that “few groups of men and women in the world need more, better, or more varied information than the 535 elected Representatives and Senators. Congress’ committees and subcommittees and members need push-button, preferably display type access to specialized banks of information. Each major bank should serve the interested committee – agriculture, appropriations, armed services, banking and currency, foreign relations, interior, etc.”
Specific examples of the use of computers in keeping track of legislation would be the Florida State Legislature which uses an RCA Spectra 70 Model 45. Here, over 5,000 bills are introduced annually and over 4,000 questions a day are asked on their status. Information is displayed on cathode ray tubes.
Having used a computer to be able to retrieve and keep track of legislation, it is also possible to use the computer to analyze the consistency of legislation. For example, it would be possible to use the computer to ensure that everywhere an idea is referred to in a bill, the same words are used to describe it. This process, incidentally, could be helpful to translators to ensure the con sistency of translations.
Needless to say, computers are already being used in a limited way to examine the consistency of treaties. For example, at Stanford University, the full text of 5,000 U.N. treaties in the U.N. treaty series, has been put on computers and experiments are in progress to see how many of these contain clauses contrary to the United Nations Charter.
No doubt, such analyses could also point out inconsistencies between Federal and Provincial legislation much in the way life insurance companies have analyzed their own policies to determine where clauses unnecessarily overlap.
One mundane way in which computers can assist the legislative process follows from the legislative production control system. Obviously, many changes are made to bills in the course of their consideration and a great deal of time can be wasted in retyping and proofreading. The Oregon State Legislature, which has one of the more advanced systems in the United States, claims that 60% of the typing and proofreading time has been saved by automating the retrieval, alteration and reprinting of legislation.
Needless to say, many of the current legal information retrieval systems have as their end product an automatic typesetting function. Characteristic of this approach would be the Linotron used in the U.S. Government Printing Office. Here a typical 8″ x 10.5″ page can be typeset in five or six seconds directly from a magnetic tape containing the latest draft of the item in question.
â€¢ Here computers assist legislators through production control techniques similar to those used in a manufacturing operation and through the use of the automatic typesetting functions now commonly used by newspapers.
5. Implementing the Law
This field is so broad that one can do little more than reference some of the major areas. For example, the law is often implemented directly by a computer. An example of this would be the computerization of income tax assessments, as carried out by the Department of National Revenue, and the IRS.
Other speakers are going to address such uses of computers as the search for legal precedents, and the retrieval of land records.
It should be noted, however, that a corollary of this type of work is that computers can be used in a wide variety of fields such as title searches, patent investigations, trademark research, etc. Many of these fields are in their infancy despite a great deal of work expended on them. For example, well defined areas of patent searching can be computerized fairly readily. An example of this could be the computerizing of searches in the steroid chemical field, where the chemical bonds can be fairly easily defined and related products determined. In other areas, the computer may, at best be able to assist in suggesting items that might otherwise have been overlooked.
In the areas of trademarks, the barrier to good computer utilization will be in the classification of shapes, colours, and sizes. Here again, pattern recognition may ultimately assist in this area.
An application such as company name searching could be greatly assisted by computers, although more sophisticated techniques would involve a SOUNDEX or AUTOPIC approach to ensure that names were retrieved that sounded similar as well as appeared similar.
â€¢ The techniques in this area involve the computer’s ability to match data.
6. Assisting the Courts
Once again, the technique of computer simulation can come to the assistance of the legal profession in one of its major current problems. In the Daily News Record of New York, June 17, 1969, it was noted that computers are now being used to clear the jam-ups in U.S. courts. Some long-held theories of what slows the court system have been shown to be untrue.
For example, in 1959, 62,000 cases were disposed of by the Federal Courts. In 1964, there had been 63 new judges added, i.e. about a 25% increase in judge-power. However, the case termination rate had increased by only 3%.
It had always been assumed that the turtle-like movement of courts was a result of a lack of judges, court personnel, or too few courtrooms.
The General Purpose Systems Simulator (GPSS) was used to establish a model of the criminal justice system. This model could:
1. Identify where the systems costs are expended, i.e. it provided a cost accounting by type of crime.
2. Provide an estimate of workloads and personnel and facility requirements based on projections of future crime and arrest rates.
3. Test the consequences of reallocation of resources within the system, i.e. by putting more resources so the courts can speed trial, the costs of pre-trial detention can be reduced. (Naturally, there are other side benefits to this.)
4. The model permitted an examination of crimes committed and suggested solutions.
The model quickly disclosed bottlenecks in the U.S. system. For example, it showed that the U.S. Commissioner was being used in only 25% of the cases allocated for preliminary investigations. In contrast, the alternative routine for preliminary processing, the U.S. branch of the Court of General Sessions was being used to 90% of its capacity, thus forming a bottleneck.
The run also suggested that a bottleneck existed in the Gran Jury section and an increase in capacity in this process would eliminate some three weeks’ delay in many cases.
It is interesting to note the type of information put into such a model. The actual data came from the District of Columbia Adult Felony cases. In 1965, there were 6,300 persons arrested in D.C. on a felony charge. Two-thirds of these did not reach the stage of a formal felony charge. Instead, the charge was reduced to a misdemeanor or the prosecution simply dropped the case. About 2,000 were held for Grand Jury action. The Grand Jury voted an indictment for 80% of these cases. The 1,600 adults were arraigned in a district court on a formal felony charge. In trial court, approximately 55% pleaded guilty, 15% were dismissed and the remainder (less than 500 went to trial). Measured from the time of initial presentment, 50% of the trial cases took more than six months to disposition and even in the non-trial cases, 50% took more than five months.
This was a very simple model but with that type of information, a model could be constructed showing the stages and delays in the process.
In general, the computer-based cost effectiveness analysis of our system of justice can help in both improving the effectiveness of policy work and court work while reducing the cost. For example, a study in Los Angeles showed the improbability of apprehension of a criminal (less than 12%) when the suspect is neither apprehended at the scene of the crime nor known to the victim. This indicated that much police work was being wasted where there was almost no chance of capture in certain types of crimes.
â€¢ Once again, simulation on computers can greatly assist the legal profession.
I have tried to review many of the computer techniques that can be applied to the field of justice. It would be incorrect to imply that the development in all of these fields had reached the same level. However, two conclusions can be drawn from this review.
1. Computers are logic machines. The law is based on logic. Therefore, computers should find a place in every aspect of legal work.
2. Lawyers know their own problems and can undoubtedly suggest many other areas where computers could assist in their work. We, in the computer field, are most anxious to expand the frontiers of this new field of legal data processing.
It has often been noted that the computer is essentially an extension of man’s mind to cope with the vast storage, retrieval, and logical operations necessary in a complex world. The use of computers in the legal field will only progress with the active participation of the legal minds that computers are intended to extend.