Harvard professor Viktor Mayer-Schönberger urges the adoption of laws requiring computers to learn to forget.  In his article “Useful Void; The Art of Forgetting in the Age of Ubiquitous Computing” available here, he warns that the burden of perpetual and persistent digital memories may increasingly constrain our willingness to engage in open society.  He appears to be particularly concerned with the accumulation of personally identifying information that, by default, is retained rather than erased:

“With the advent of ubiquitous computing, of cheap GPS chips in our cell phones, cameras and cars, of RFID tags in everyday objects, and of tiny, networked sensors that surround us, a more comprehensive trail of our actions will be collected than ever before. Given low cost of storage, ease of retrieval and potential value in accessing information, much of the data that is being collected will be kept for months if not years, as our societal default has shifted from deletion to retention.”

His solution?  A combination of law and software which would cause data, upon reaching a set expiry date, to be automatically deleted, unless a human specifically chose retention instead.  The goal of the proposal is to “reinstate the default of forgetting.”  However, while it is true that the default for human memory is forgetfulness, the default for digital information is also forgetfulness, just on a non-human scale; entropy will, sooner or later, destroy all information, whether encoded in neurons or bits.  Mayer-Schönberger’s proposal would just speed up the process for digital memories, thus reducing the able memory of a computer to that of a human.  This is unfortunate, because for all of recorded history we know that humans have been striving to overcome those limits.

But, the pervasiveness and longevity of digital data are hugely beneficial in ways that Mayer-Schönberger doesn’t acknowledge, including: the historical value in preserving even the most mundane records, the way computers lovingly tend to storing and remembering the most dreadfully boring facts, or their ability to manipulate vast human behavior data sets on the individual or collective level.

Fortunately, Professor Mayer-Schönberger’s proposals stand almost no chance of being adopted.  Aside from sunsetting provisions requiring the destruction of certain government records–yes, it’s scary when governments have too much information, a fact Mayer-Schönberger acknowledges when he compares the Soviet Union’s information-retention policies with those of Google. Most of us have decided that the advantages of having such extensive electronic records outweigh the downsides.

But if there’s any doubt, here’s a test: start a new commercial search engine that doesn’t maintain search records, or one that periodically wipes all records, and see how well it performs in the free market.  Search engines like that used to exist, but they were inferior and ultimately rejected by consumers in favor of modern engines that incorporate data from past searches to fine-tune results on an ongoing basis.

Someday we will be gone, all of us.  Only our data will survive.  Our computers and their descendants will continue their silent, eternal vigil over our digital pictues, e-mails, blog posts, and yes, cell phone calls and receipts.  Those things are part of our legacy and will provide a rich resource for future historians.  For both our own benefit as well as for posterity, we should seek to improve data retention, not cripple it through unnecessary laws.