Monday, August 21, 2006

Computing news reflects odd interests

Are the machines learning from us? (Or, to recycle a well-used Bushism, “Is our computers learning?”) In times past, it was panicky, circulation-hungry editors and vacuous TV producers who gave voluminous and unending coverage to the most meaningless news items of their day. Remember Michael Jackson, not so long ago? He’'s famous, as well as bizarre, so his second child-molestation trial somehow warranted massive amounts of column inches and airtime. Remember the OJ trial? It got so much attention the media had to invent reasons why it was significant. (I'’ll grant that it helped illustrate the racial divide in this country, but it'’s kind of pathetic that anybody still needed illustrations of that from the trial of a previously minor celebrity.)

Now, though, Google News can help us focus on celebrity tittle-tattle without the need for any human agency. The service'’s “About Google News” page says that sophisticated algorithms are used to determine the placement of stories on the site’s main page. Incredible. Technology marches on.

Unfortunately, the site also states Google News’ choice of top stories is based on the “collective judgment of online news editors,” which probably explains why the last 10 or 12 times I'’ve logged on, there’s been some story like “"JonBenet Ramsey Suspect Picks Nose!"” right at the top of the page. As of yesterday, Google News links to 3,588 JonBenet-related items. The second top story, about nascent holy war in Iraq, links to 795. Apparently, the whole affair wasn'’t embarrassing enough for American journalism the first time around.

Clearly, the machines have learned much from us.

Jonathan Gold

No comments: