Intellectual Property Law

The Curious Case of Google's Out-of-Date Articles

By Cynthia Brumfield

A seeming tempest-in-a-teapot brewed up this week when a 2002 article on United Airlines (UAL) filing for bankruptcy on a Tribune Co. website got scooped up by Google's automated search software and was mistakenly presented as "new" on web news pages. The "new" information promptly sent UAL's stock plummeting to almost nothing before investors got wise.

Tribune blames Google for the problem, saying that it has asked the search giant to stop conducting searches for stories on its websites -- Google says that Tribune asked it for no such thing. This snafu has been elevated to a level of an SEC investigation with the idea that someone may be on the hook for spreading false information about publicly traded securities.

It's hard to say exactly what happened here. Google says that its Google News' webcrawler picked up the article because it appeared on Tribune's Sun-Sentinel website with a "fresh" date on the page and the bot thought it was a new article, particularly given that, according to Google, the article had a new URL. As this forensic analysis suggests, in Google's world, if something has a brand new URL, then it's a brand new page. Tribune denies the article had a new URL.

Whatever the case may be, I've discovered that Google News, at least so far as Google Finance is concerned, will often display "old" news. I discovered this over the past few months as I've pursued my new (and hopefully temporary) obsession with stock market investing.

Indeed, I've noticed that Google Finance will sporadically display news that is months old. But because I assumed that there were system glitches, and because I've been tracking financial news extremely closely, I understood the news to be old and never mistook it for recent information.

And, because I never anticipated the kind of reaction generated by the old UAL story, I didn't perform screen captures to document Google's intermittent display of old news articles. However, this morning I did. Here's a screen capture I made at 9:47 a.m this morning and as you can see, Google Finance displayed articles that were from September 10 and September 11. Today is September 12.

The lead article from Forbes is (as is clearly indicated on the Google Finance page) 22 hours old and deals with a weak opening to the stock market (weirdly enough, the same thing happened today so folks not paying attention to the date might get confused).

Another article from Forbes is from September 10 and discusses how Texas Instruments' earnings report boosted after-hours trading in the tech sector. Another article is from the Market Oracle discussing an intraday double-bottom for the S&P 500 Depository Receipts -- dated September 10.

I'm going to start making screen captures whenever I see these weird, seemingly glitchy old news articles. They may or may not have anything to do with how the Googlebot retrieves articles or even be related to the UAL incident.

Still, it's pretty odd how often it happens.

Full post as published by IP Democracy on September 12, 2008 (boomark / email).