Here’s a fascinating dispatch from the new world of reputation management. The New York Times is now apparently receiving one request a day from people who want the paper to remove an old article from their online archive — because the article contains incorrect or incomplete information that makes the person look bad, and it’s cropping up on Google.

It’s an incredibly fascinating and troubling issue. The Times, like most newspapers, often runs a news brief when someone gets in trouble — but doesn’t print a followup when they’re cleared of their allegations, because it seems less newsworthy. In the past, this caused the subjects a lot of heartache, of course. But it’s far worse now, because a prospective employer, business partner or spouse Googles the person and … whoops, there’s the original news item, in the #1 or #2 slot on Google, still uncorrected, decades later.

What’s the answer here? Clark Hoyt, the Times’ public editor (pictured above) tackled this one in his weekly column today, and discovered that the news editors at the paper are baffled about what to do. They could acquiesce and remove the articles, but they’d worry about where to draw the line; they don’t have enough resources to re-investigate every two-decade-old article that subjects complain about. (Someone could, of course, complain about a perfectly legitimate article in hopes of having it taken down.) And if they start changing or removing old articles, it could begin to erode the trust of those who use their archives for research: “What’s not here now, and why isn’t it here?”, they’d start to wonder.

The most interesting suggestion came towards the end of the column:

Viktor Mayer-Schönberger, an associate professor of public policy at Harvard’s John F. Kennedy School of Government, has a different answer to the problem: He thinks newspapers, including The Times, should program their archives to “forget” some information, just as humans do. Through the ages, humans have generally remembered the important stuff and forgotten the trivial, he said. The computer age has turned that upside down. Now, everything lasts forever, whether it is insignificant or important, ancient or recent, complete or overtaken by events.

Following Mayer-Schönberger’s logic, The Times could program some items, like news briefs, which generate a surprising number of the complaints, to expire, at least for wide public access, in a relatively short time. Articles of larger significance could be assigned longer lives, or last forever.

Mayer-Schönberger said his proposal is no different from what The Times used to do when it culled its clipping files of old items that no longer seemed useful. But what if something was thrown away that later turned out to be important? Meyer Berger, a legendary Times reporter, complained in the 1940s that files of Victorian-era murder cases had been tossed.

“That’s a risk you run,” Mayer-Schönberger said. “But we’ve dealt with that risk for eons.”

Programming a database to forget: I love it! This whole issue is another symptom of our increasingly weird digital world, where feats of memory that are superhuman — or inhuman, or both — are made possible via silicon. Last year, when I profiled Gordon Bell, the Microsoft researcher who’s trying to record every aspect of his daily activities in a “MyLifeBits” data, it raised a lot of deeply personal questions about the relative value of remembering versus forgetting. We humans rely on our faulty memories to make sense of the world, because remembering everything would drive us nuts; one definition of “wisdom” is “all the knowledge that’s left over after you’ve forgotten the less-important things you’ve ever learned”. But of course, having perfect recall can allow for new and deeply cool forms of knowledge: Google’s great at tying together strands of information I wasn’t even aware were connected until I hit “search”.

Has anyone ever tried to do what Mayer-Schönberger suggests — and model the act of forgetting in a database? In a way, you could argue that Google sort of already does this … insofar as any piece of data appears on page 57 of a search query is essentially forgotten from the overmind, because almost no-one will ever read it. By this logic, one of the best ways to try and get Google to “forget” you is to seed the Net with really high-quality pages about you, which Google will find ever more interesting, driving the undesirable stuff downwards. This is sort of what I’ve always argued people should do if they’re unhappy with their Google identity: Start blogging, because it’s a pretty sure-fire way of eventually dominating the #1 slot, if you work hard and become good at it. Even so, though, there are no guarantees that the old Times report about your unjust drunk-driving arrest won’t appear someone on the first page. There’s no silver bullet here.

This is an issue we’re going to hear and more about in the future, I predict.