It's Google. This Must Be Belgium.
As cyber-geeks the world over know, last week a Belgian appeals court upheld the order of a Belgian trial court forbidding Google News from including Belgian newspaper content on its website. (I have been unable to locate an English-language version of the most recent decision. The link above includes an English version of last year's trial court ruling.) The latest decision includes fines of a few million dollars for the time during which Google was unlawfully linking to the Belgian news sites, which is real money to normal people but not to Google. I admit to not following this story quite as closely as some of my more teched-up friends, and thus I could be getting this all wrong, but with that disclaimer, I must say I don't have much sympathy for Google's argument in this dispute.
Google News aggregates news stories from newspapers around the world and displays them on a single web page. Clicking on a given story directs you to the website of the newspaper from which the story is drawn. There are 4 potential copyright violations here, and based on the terse trial court opinion, it's hard to tell whether Google was found liable for all of them, but it's worth noting them in any event.
1) Deep-linking. Some European courts have found that so-called "deep linking," in which one site links to sub-pages rather than homepages of other sites, is an unlawful appropriation of the linked site's intellectual property. Google News clearly deep links. I don't know much (read "anything") about Belgian copyright law, but as a general matter, it strikes me as unwise for copyright law to forbid deep-linking, which is essential to much of the work of internet search engines.
2) Copyright in blurbs. The very short descriptions of news stories that appear on the Google News homepage may contain copyrighted info, at least for "lead stories," which typically contain the first one or two sentences of the linked story. I think copyright law ought to treat this as fair use, but perhaps Belgian law doesn't.
3) Robotic copying. In order to generate the content for Google News, Google's own servers need to make electronic copies. That act of copying may itself be deemed illegal. This is similar to a provision of the Digital Millennium Copyright Act in the U.S., which makes the playing of a legally purchased CD or DVD by a computer technically illegal, because the copyrighted code is first loaded into RAM, i.e., copied. This is just stupid, in both contexts. Whether via fair use doctrine or some doctrine of implied license, this shouldn't count as a violation, so long as a new copy is not made available to others.
4) Cache. Many of the Belgian newspaper plaintiffs have a policy of posting their content free on their websites for a limited period and then making it unavailable or charging for it after that. However, Google News makes available a free cached copy of material on its own website even after the newspapers have taken the free material down. I don't see how that isn't a blatant copyright violation. Case closed.
To be sure, Google could rationally have a policy of only posting links to a news source on Google News if that source agrees thereafter that Google News will be able to post the story via cache even after the source removes it from its own free site. But it hardly follows that Google gets to impose that policy on Belgian (or other) news sources without their consent, or that Belgian (or other) news sources should be presumed to consent unless they include lines of code in their pages that block Google from aggregating. I think I agree with those observers who say that the newspapers have more to gain than to lose from inclusion in Google News, but the point here is that it should be their choice to make.
Unless I've misunderstood this whole case. If so, never mind.
Google News aggregates news stories from newspapers around the world and displays them on a single web page. Clicking on a given story directs you to the website of the newspaper from which the story is drawn. There are 4 potential copyright violations here, and based on the terse trial court opinion, it's hard to tell whether Google was found liable for all of them, but it's worth noting them in any event.
1) Deep-linking. Some European courts have found that so-called "deep linking," in which one site links to sub-pages rather than homepages of other sites, is an unlawful appropriation of the linked site's intellectual property. Google News clearly deep links. I don't know much (read "anything") about Belgian copyright law, but as a general matter, it strikes me as unwise for copyright law to forbid deep-linking, which is essential to much of the work of internet search engines.
2) Copyright in blurbs. The very short descriptions of news stories that appear on the Google News homepage may contain copyrighted info, at least for "lead stories," which typically contain the first one or two sentences of the linked story. I think copyright law ought to treat this as fair use, but perhaps Belgian law doesn't.
3) Robotic copying. In order to generate the content for Google News, Google's own servers need to make electronic copies. That act of copying may itself be deemed illegal. This is similar to a provision of the Digital Millennium Copyright Act in the U.S., which makes the playing of a legally purchased CD or DVD by a computer technically illegal, because the copyrighted code is first loaded into RAM, i.e., copied. This is just stupid, in both contexts. Whether via fair use doctrine or some doctrine of implied license, this shouldn't count as a violation, so long as a new copy is not made available to others.
4) Cache. Many of the Belgian newspaper plaintiffs have a policy of posting their content free on their websites for a limited period and then making it unavailable or charging for it after that. However, Google News makes available a free cached copy of material on its own website even after the newspapers have taken the free material down. I don't see how that isn't a blatant copyright violation. Case closed.
To be sure, Google could rationally have a policy of only posting links to a news source on Google News if that source agrees thereafter that Google News will be able to post the story via cache even after the source removes it from its own free site. But it hardly follows that Google gets to impose that policy on Belgian (or other) news sources without their consent, or that Belgian (or other) news sources should be presumed to consent unless they include lines of code in their pages that block Google from aggregating. I think I agree with those observers who say that the newspapers have more to gain than to lose from inclusion in Google News, but the point here is that it should be their choice to make.
Unless I've misunderstood this whole case. If so, never mind.
posted by Michael C. Dorf @ 12:08 AM
7 Comments:
At 6:33 AM,
KipEsquire said…
Regarding #4: It takes only one simple line of HTML code to instruct search engines not to cache, just as it takes only simple line of HTML code to to tell search engines not to index a site in the first place.
So if you believe in the doctrine of the low-cost avoider, who's problem is this: Google's or the Beligians?
At 9:21 AM,
Michael C. Dorf said…
That's true from a pure law-and-econ perspective, but, as some other bloggers have noted, by this logic, we wouldn't prosecute burglars who enter through unlocked doors, because a homeowner who didn't want burglars to enter his home could simply lock his door.
At 9:31 AM,
Caleb said…
Lurking alongside the cache problem is the problem that competitors apparently pay the Belgian news sources to do what Google is doing for free (have a directory of the stories and linking them). In that case, they may be looking to prove copyright in the parts of their stories that google reproduces on its search page so that they can get injunctive relief and force Google to start paying as well. (I should note that I'm indebted to Prof. Ginsburg for this information, and not to any special knowledge of Belgian papers)
At 11:07 AM,
egarber said…
This post has been removed by the author.
At 1:29 PM,
egarber said…
To Mike's point -- just thinking about how an analogy could work the other way.
Maybe this type of "opting out" is similar to a scenario where personal information -- addresses, phone numbers -- is sold to telemarketers. Even though "property" of some sort (though I readily acknowledge a phone number isn't the equivalent of copyrighted material) is used for commercial purposes by a third party in this scenario, the burden is on the resident to "opt out", either with technology or via the federal no-call list. In other words, maybe posting material on the internet is the rough equivalent of giving up some degree of personal information when applying for a credit card, etc.
I'm pretty sure Google won somewhere in a previous case on this type of estoppel argument.
At 2:37 PM,
kb2188 said…
Google has won several cases, but mostly against online publishers of images. These cases turn on a fair use argument, that the thumbnails Google provides are different from the originals and cannot supplant them. That would probably not work, for newspaper articles, if the Belgian case were to be heard under US law.
I would be surprised if Google succeeded on that estoppel theory. That amounts to requiring a copyright owner to comply with certain formalities. The rest of the world has opposed formalities in copyright for at least a century, the US for a couple decades.
Moreover, most of the US copyright statute represents Congress's fine-scale tinkering with copyright entitlements. In the face of sch regular Congressional activity, it would be surprising for a court to undertake its own adjustment.
At 6:41 PM,
egarber said…
to kb2188
In looking around, I think it was a district court victory for Google.
Here's a blurb from a Stanford website:
http://cyberlaw.stanford.edu/packets003808.shtml
"Next, the Court considered the defense of estoppel. Estoppel has four elements: the Plaintiff knew of the infringing conduct, Plaintiff intended Defendant to rely on his conduct, Defendant was ignorant of the true facts, and Defendant detrimentally relied on Plaintiff’s conduct. The court found all four elements were met. Plaintiff admitted he knew that without the no-archive meta-tag, Defendant would provide a copy of his work to users through its cache. Plaintiff chose not to include the no-archive meta-tag knowing that Defendant would rely on it in building its cache. Defendant was unaware that Plaintiff did not want his work cached. Without Defendant’s reliance, there would have been no lawsuit. Therefore, the Court held that Plaintiff was estopped from his claims as a matter of law.
Post a Comment
<< Home