One of the seeming paradoxes of the networked communications revolution is how converting information from tangible, physical form to streams of electronic bits makes that information less, rather than more, transitory and ephemeral. There are plenty of factors we might cite as contributing causes — the cost of mass storage is now tiny and still shrinking, connectivity (at least in the developed world) is increasingly robust, and communications architectures are springing up to take advantage of these capabilities. Letting information live “on the network,” as it were, as distinct from on the hard drive of any individual PC, extends its longevity to the theoretically infinite.
In some respects, this is an exciting and positive development. I’ve become a big user of Web 2.0-type services and online document storage. All my current photos live on Flickr, and I’m rapidly uploading my earlier archives. My e-mail (after scrubbing by Spamcop) ends up on Gmail. My university provides me with SSH-accessible file storage for my documents, and as the kinks continue to be worked out, I’m sure I’ll begin to rely more and more on services like Google Docs (née Writely) to create and manage my word-processing and spreadsheet output. Particularly for someone like me who owns and uses multiple computers, having my “workspace” preserved online, accessible from anywhere I can get a network connection, is quite liberating — it means I no longer have to worry about which computer has the most current version of a file I’ve been editing, and it saves me from buggy hardware. This, I think, is the kernel of truth at the heart of Sun’s old axiom, “the network is the computer.”
But the permanence of networked information has costs, too, which (like the benefits) are only beginning to be explored. Members of the generation just behind mine, who have grown up reflexively creating and posting information online, are learning that digital is forever — if you’re a job applicant (or even a camp counselor), anything that has ever been written by (or about) you online is, at least potentially, still there. (Back in my day, we used goofy aliases to hide our online identities; but I gather that practice has been fading.) Once information is online, it turns out, it may becomes quite hard ever to get it back offline again — the Wayback Machine preserves old web pages; Google Groups archives Usenet posts; and it’s only a matter of time before somebody comes up with the magic bullet that automatically archives IRC and IM conversations and makes them searchable. Even your deleted e-mails aren’t necessarily gone; they may still exist on backup tapes where law enforcement authorities can get them. The durability of digital content raises problems that touch on both informational security and individual privacy.
It’s in that context, as yet another illustration of the new problems of information persistence, that I found today’s NYT article, Expunged Criminal Records Live to Tell Tales (registration required, unless) so interesting. (Dan Solove blogs about the story here). This is one of those rare situations in which the legal system has a strong interest in the affirmative concealment of truthful, factual information — in this case, information about the criminal record of those convicted of (usually relatively minor) transgressions. Although that information can be legally ordered expunged or placed under seal, once the information escapes the control of the court that issued the original judgment, any subsequent redactions of the record might not be reflected (in much the way that Google doesn’t go back to its old backup tapes to remove e-mails that you deleted after the backup was created). The NYT fingers companies that purchase the judicial records for use in compiling commercial databases as the bad actors here, and Professor Solove proposes contract terms that would require the database companies to maintain synchronization between their records and the courts’. But the broader social issue — information, once released into “the wild,” tends to resist control — transcends this particular problem and can be expected to pose recurring headaches for policymakers in the future.