The paper below was written by me in the Nordic magazine Data No. 3, 1974. It phrophesis much of what Internet has become today. The dreams were with us also 23 years ago, even if the dreams would not be realized until today.
The use of computers may increase hundredfold in the next two decades. This will mean a revolution to society. It can give a more authoritarian society, but it can also give a more democratic society. In the democratic version, computer terminals at public libraries may give anyone access to all facts in the government data banks (with some special exceptions).
We must inform the politicians of these facts, since decisions taken today are necessary to guide the evolution of computers in the right direction.
2. The future use of computers
3. Social and political impact
3.1 The black scenario: The authoritarian use of computers
3.2 The white scenario: The democratic use of computers
4. The democratic information system
5. Is this technically and economically feasible
6. How to step in the right direction
During the last years, there has been a large amount of discussion about the social and political implications of computers. But much of this discussion has been narrowly focused on the problem of personal integrity and the invasion of privacy. In my opinion, this special matter must be viewed as part of the much larger question of computers and society. Without this wider viewpoint, the political decision makers may arrive at conclusions giving something quite different from what they intended.
To consider the political effects, we must first see which possible uses computers can have in the future. We must also see which alternative choices are available, to be able to guide the future in the right direction.
Technical prognosis makers claim that computers will rapidly get much cheaper in the future. Computers will soon cost only a tenth or perhaps even only a hundredth of what they cost today.
This decrease in price will mean an increase in the use of computers, so that we can expect that the total amount of computer power available may be fifty or a hundred times as'large as today in ten or twenty years.
The increase will however not mean that all applications of today will increase with the same rate. Some applications give a large gain from the use of computers. These applications are already exploited near to their limit. Other applications give a smaller gain from the use of computers. Those applications are often not profitable today, and are therefore only done on a small, experimental basis.
With lower computer prices, the large increase will be in these applications with smaller gain than todays applications.
One such application which probably will increase very much is the use of computers to store, collect and disseminate documents. By documents I mean e.g. reports, letters, announcements, investigations, catalogues, offers, laws, legal decisions, contracts, news items, literary word etc, etc.
Such documents do not have as much formal structure as conventional computerized data bases. Computers are most efficient when the data handled has a formal structure. Computer handling of documents is therefore not yet always profitable, but will be so with cheaper computers in the future.
A typical office worker in the future will have his own display terminal at his desk. He will write most of his memos, notes, messages, letters etc. on this terminal. When he wants someone else to look at a document, he will just store a command into the computer, and the receiver will, when he is free, be notified by his computer terminal that there is information for him in the computer.
The sender can also disseminate his document to several other people simultaneously. The receiver can send an answer to the sender, or they can add a comment to the document or suggest a change in a new document, appended to the previous one in the computer.
Updating of documents will be very easy. Many documents can be copied directly from a standard form which is slightly modified.
Instead of having the sender decide who shall see a document, the sender can add search keywords to the document. Anyone asking for documents with those search keywords will then find it. Thus, the receiver and not the sender can decide to whom the document shall be sent.
A variation of this is a user who has a permanent "interest key" stored in the computer. As soon as a new document arrives which fits his "interest key", he will be notified. This also means that the receiver, not the sender, governs the distribution of the document.
The dominating use of computers for document handling in the future will mean a revolution to society, much more thoroughgoing than the changes arising from e.g. nuclear power.
To understand the choices which we can make, I will draw two scenarios representing extreme alternatives:
The use of computer is in the hands of a few large institutions like government agencies, large private companies or large trade unions. These institutions decide what shall be done and not be done with the computers. They use computers to increase their own power.
The ability to get more information faster is available only to these institutions (the so-called "need-to-know" principle). Therefore, they have all the facts and thereby also get all the power. Ordinary people are more or less helpless against these large institutions with their power gained by access to facts.
The dissemination of information is also governed by these large institutions. They decide who shall get facts and who shall not get facts, They use the power of the computer to select just those people who can be influenced in a certain direction by certain facts, in a way which is good for these large institutions.
This is only to go a little further in the direction where mail advertising is moving today. Computers are used more and more to penetrate the defenses of the people to whom the advertising is sent. For each individual, the computer can select the arguments to persuade him to do what the advertiser wants him to do.
The facts stored in the computers are available to anyone. Anyone can go into a public library, sit down at a terminal and ask questions to the computers. All government documents (with some special exceptions) and many private company documents are stored in the computer and thus available to everyone.
All people can also have an interest keys stored in the computer, which is used to govern what information is sent to them.
Examples of questions which someone might want to ask at one of the public terminals:
Example of an interest profile, which someone might store in the computer: "(Stockholm and pollution and industry) or (food and health)".
The advantages with this system as compared to the authoritarian system are:
The general democratic public information system can be seen as an alternative to the so-called "management information systems". These latter systems are intended to increase the power of managers by giving them all facts. The democratic system is intended to increase the power of the people by giving them access to all the facts.
All non-secret government documents which are stored in computers should also automatically by law be transferred to the democratic computer system.
Many documents in large private companies and institutions should also by law be transferred to the democratic system.
Anyone can for a certain small fee store his own document in the general system, and give it a search key so that other people can find it. He can either give it away free, or he can ask the system to charge those who access the document. The income from the charges are transferred to the copyright holder.
This will work in many ways similar to a newspaper. But in a newspaper, the editor decides what to publish and not publish. Most newspapers throw away 90 % of everything they get. Thus, the information in a newspaper has always been censored by the editor. In the democratic system, the reader does the censoring himself. All the facts are in the computer, and the reader decides in a dialogue with the computer what to read.
He might sit down at a terminal and ask an initial question. The computer may tell him that there are 2567 documents relating to the question. He can then narrow down his question in various ways e.g. by rephrasing, by setting a creation time limit etc.
He can also say that he only wants documents from certain institutions or documents produced by certain authors.
A group of people with similar opinions can also form societies with the purpose of selecting information. These societies go through the material in the computer and gives each document a judgment: Very interesting, Moderately interesting, Faulty but interesting, Not interesting etc. A member of a society can then at a terminal ask for only documents which have got good marks from his society. In that way, a group of people can cooperate to help each other finding the important and interesting facts in the mass of unimportant documents.
It there is a charge on reading a copyrighted document, the receiver is asked if he accepts the charge before he finally gets to read the document.
The importance of this system are basically two factors:
We do not know yet if it is technically feasible, but man has created other technical systems, for example the public telephone network, of very large complexity.
Economically, the democratic information system is not possible with today's computer costs. But if computers cost a tenth or a hundredth of the cost today, then the system may cost less than what we pay today for a daily newspaper, and this will probably be acceptable.
Decisions taken today can guide the future in the right direction.
First, the general principle of general public access to documents must be made into law, if it is not yet there. This law must be upheld even for computerized documents.
Obvious exceptions are necessary, e.g. for certain individual personal data and for military secrets. But these exceptions should be kept as small as possible.
Misuse should be countered not by increased secrecy, but by laws governing the use of computers. An example: The misuse of computers to produce certain kinds of mailing advertising should be countered by a law which says that mailing advertisement should not be sent to people who have informed the address register that they do not want that kind of advertising. Computerized mail address registers must thus contain space for interest profiles, and these interest profiles should be set by the receiver and should be used by the advertiser. In that way we do not need to keep as much personal information secret as we otherwise would have to.
All government data banks should be programmed in such a way that they can answer one general simple question: The question consisting of a search key with and and or connectors. Example:
"Stockholm and industry and (pollution or sewage or garbage)". Such a search key should search all the fields in the documents, if the questioner so requires. If the fields are coded, they should be translated into natural human language before the search.
When someone asks for a print of a document in a government data bank, he should get the whole document printed in a legible form. Codes should be translated into natural language.
Research on document data bases and on natural language communication is also important. If we want everyone to have access to computers, then the communication must be in a simple language, preferably natural language or a subset of natural language. Research on the psychology of the communication situation between humans and computers is also important.
Decisions made today will influence whether the future use of computers will give a more democratic or a more authoritarian society. If we only blindly look at the risk for "invasion of privacy" we may inadvertently push towards the authoritarian future. We may later regret such decisions.
We must inform the politicians of the whole picture of future computer usage, not only the "invasion of privacy" bit. Otherwise we may misinform the politicians into making dangerous decisions.