I have started to work a bit on a forthcoming paper on the history and conceptual thrust of probabilistic indexing in Information Retrieval (“naive Bayes classifiers” for the connoisseurs) – which will also be a chapter of a forthcoming book – and while researching I stumbled over a beautiful paper. E. M. Maron, one of the central pioneers in that field, worked at the RAND corporation in the early 1960 and when not developing prototype systems that used Bayes’ theorem for “fuzzy” document scoring, he wrote memoranda on wider subjects, such as cybernetics ; “Computers and our future” is a very short piece from 1966 that is extremely lucid in terms of the questions it asks. Consider these three points about “the basic characteristics of machines [computers] and at their implications”:
- Computers operate at exceedingly high speeds. What does this imply? This means that if a high speed machine is used to control a complex situation, then it could compute an action to be taken and execute that action before a human could intervene. What are the potential dangers?
- Computers, at least at present, demand extreme precision in their instructions. They take their instructions literally. Could there be a tendency to delegate a complex decision to a machine and find out that the machine did what we asked, but that it was not what we wanted – because we ourselves did not fully comprehend the fine structure of our own instructions? What are the full implications of this?
- Computers have the capacity to handle large amounts of data. They can digest, analyze and relate these data in complete detail. If these data concern financial and personal information on people, what are the implications for the concept of privacy, for improper manipulation and control? What happens when large amounts of information about the economic and political aspects of a society are fed back to the citizens of that society? What is the influence of this information on the truth of the information? Can this type of information, when fed back, cause instabilities – economic and political? (Could the information flow – feedback, overload, and instability – be modeled?)
I would argue that these are extremely timely questions, and they show how aware at least some of the technical pioneers were in terms of the wider implications of the work they were doing. There is a tendency to caricature the cyberneticians of the 1950s and 1960s as narrow-minded taylorist technocrats, but I guess the story is more complicated after all…
Tech support questions will not be answered. Please refer to the FAQ of the tool.