In the past the storage and recall of information (the act of remembering) was limited. If people wanted to keep a record, it had to be written down (at great expense in the days before printing) or they had to rely on (notoriously error-prone) human memory. As time moved on, more and more could be recorded, but recall in the analogue world remained difficult - the raison d’être of information science. However, with the proliferation of digital recording and the advent of cheap and vast storage, the balance has shifted. In a world deluged with data - including our personal collections of digital photos, email inboxes and the like - it has become easier to record everything than attempt any kind of manual pruning - deleting bad photographs, irrelevant emails, etc. At the same time recall methods have advanced, so that I can (should I want) look up an email I sent more than five years ago or see what a colleague blogged last year.
This shift from people forgetting to machines remembering is the central theme of Viktor Mayer-Schönberger’s book Delete, published by Princeton University Press. And a fascinating, frightening, well-argued and accessible read it is!
The work opens with the now familiar horror stories of digital remembering - a newly qualified teacher failing to get a job on account of the picture of her on a social networking site, the psychotherapist refused entry to the USA because of a (presumably open access) journal article, published several years before, in which he mentions having taken LSD in the 1960s. Examples of how society is now able to discover (to remember) facts about your life that you have forgotten yourself. This opening chapter neatly sets the scene and highlights just how much Mayer-Schönberger has read around the problem of ‘perfect memory’.
In the next two chapters, the reader is taken on a ride through the psychology, sociology and history of forgetting and humankind’s battle against it. They culminate in the rise of the technologies that now leave us with a society capable of seemingly perfect memory (though not everything is remembered digitally and we’d do well to remember that!) while individuals are now capable of looking up their past in ways hitherto unimagined. Though the author’s arguments are compelling, they are on occasion overstated or based on false premise - is it really all that easy to mine the vast data resources we have at our command yet? However, the book does get the reader thinking and the author himself states that part of his reason for writing it was to stimulate debate.
Chapter IV, Of Power and Time - Consequences of the Demise of Forgetting, is probably my favourite - perhaps because I love a good tale of doom! It examines the consequences of total recall and boils the issue down to two fundamentals. Firstly there is the loss of power, as information about us is duplicated and reused (often out of context) with or without our permission; and secondly the negation of time. Among the issues are two terrifying possibilities: that perfect memory threatens reason, abstract thought and the ability to make decisions in the present, and; that the reasons for retaining data now may seem sensible, but what if (as with the chilling example given) it should fall into the wrong hands in the future?
Having clearly and concisely built a picture of the problem, from the early days of human history to the present, Mayer-Schönberger then turns to some potential solutions. Chapter V outlines six potential responses, drawing on information privacy issues as well as other areas. Curious here is the way he suggests one response would be for information sources (us) to use digital rights management techniques to ensure our data are safe, effectively turning the tables on the music industry or search engines. It is an interesting idea, though one the author later dismisses.
The final chapter before his conclusions, entitled Reintroducing Forgetting, goes on to give Mayer-Schönberger’s own solution to the problem - data with a shelf life. That is to say, every piece of information - from files on your hard drive, through your emails and even your search terms and online purchase history - is given an ‘expiration date’ (by the end of this chapter, try not to scream at the number of times those two words appear!).
It is here that I find my only real criticism of the book. Up to this point we’ve been taken in, we are willing to overlook the odd overstatement (information retrieval is easy for example, or that you really would forget a falling out you had with an old acquaintance only to rediscover the event in an email long forgotten). However, Mayer-Schönberger’s suggested response leaves a lot to be desired. Perhaps this is deliberate on his part, perhaps you will disagree with me, but it seems this chapter is not as well researched or carefully considered as the rest of the book. It seems unlikely that an end-user would specify a data expiry date - no matter how simple the interface - and the solution would not be as easy to implement as is suggested. The example of a camera that reads its subjects’ data wirelessly and adds an expiry date to the photo feels entirely unworkable!
But do not let these final thoughts put you off this book! On the contrary, I hope they provoke you to go and read it for yourself, because Delete is a fascinating read and one I would recommend to every information scientist I know. Viktor Mayer-Schönberger is right: there never has been a more pressing need to question total recall.