Could Your Website Be Liable For The Way Google's Algorithm Summarizes It?
from the that-doesn't-seem-right dept
Ruby writes in to let us know that a Dutch website, Miljoenhuizen.nl, has been found liable for the way that Google summarized the content on the website. Google, of course, has algorithms that try to summarize the contents of a page in a snippet so that you know what’s behind the link, and how it relates to the search that you do. As a part of that, it often will show parts of sentences connected by ellipses, and that’s what happened here. The snippet on Google read:
Complete name: Zwartepoorte Specialiteit: BMW…This company has been declared bankrupt, it has been acquired by the motordealer I have worked for Boat Rialto…
This upset Zwartepoorte, an auto dealer, who felt that this summary falsely stated that it had gone bankrupt… so it sued the underlying site. It’s quite surprising (on a number of different levels) that it didn’t sue Google as well (or instead). However, the court actually agreed that this was the fault of the original website owner, and told Miljoenhuizen.nl to fix the website so that Google wouldn’t summarize the site that way. It’s hard to fathom how this could possibly be Miljoenhuizen.nl’s fault, though apparently someone from Miljoenhuizen.nl suggested that it could control that in the courtroom — which likely resulted in the judge’s ruling.
Filed Under: liability, search, snippets, summaries
Companies: google
Comments on “Could Your Website Be Liable For The Way Google's Algorithm Summarizes It?”
Actually it is fairly easy to control what google summarizes. One of the ways to assure that you don’t run into this problem is not to use “This company” in the same paragraph as another company’s info.
If they had replaces “This Company” with “Bob’s Car Parts”, there would be no confusion.
It’s a weird case, but if one of the defendants admitted they could control the summary, perhaps they could have acted to avoid confusion.
Re: Re:
So if I jump through hoops, spun around three times, and do the hoky poky maybe I can get away with being sued.
Re: Re: Re:
huh?
Um
…it’s an odd bit of liability, true. But it’s also true that the site could easily have changed Google’s summary by including a tag. Google honors those and uses them as summaries. They only generate their own summaries of that tag is missing.
So yeah, the site could have changed it if they had wanted to. I’m still not seeing the liability, though. It’d be like requiring someone to change their signage because of what a newspaper wrote about it.
Shallow Pockets
For once someone that should be suing Google is suing the smaller Party without the deep pockets!
We should rejoice…and yet we just die a little more inside.
Re: Shallow Pockets
And by “should be sueing Google”, I mean that in the context of “If they were totally idiots and didn’t understand a thing about the Interwebs, and were looking for a target to blame for some random occurrence that caused them a minuscule amound of negative blowback.”
Then, totally Google.
ofcourse the original snippet was in dutch but would only apear if looking for the words “Zwartepoorte” and “bankrupt” and google puts together snippets around those two words, if you would google “techdirt” and “bankrupt” some slightly suggestive snippets come up aswell
Google Can't Be Sued
Google would be completely protected by 47 USC 230 for site descriptions, even if its algorithms select the snippets to put together. See, e.g., the Maughan v. Google case, http://blog.ericgoldman.org/archives/2006/10/google_wins_law.htm and the Murawski v. Pataki case, http://blog.ericgoldman.org/archives/2007/09/askcom_not_liab.htm
Even though a website can try to control its description through various HTML commands, the ultimate decision about which information to present is Google’s, not the individual website. http://ssrn.com/abstract=635803 So it would be ironic if a website would be liable for a description that it didn’t actually create.
Eric.
Re: Google Can't Be Sued
Only in America. Last time I checked, US law doesn’t apply in other countries.
Re: Google Can't Be Sued
Google would be completely protected by 47 USC 230 for site descriptions
Well, since the case took place in Holland, there maybe different laws. 🙂
But yes, I agree. Neither should be liable…
Mike
alternatives
The article is not entirely clear, but I got the impression that the defendant had been asked to change the site and did not. If that is the case, then what alternatives were left to the complainant ?
Was there any motive for the defendant to do this on purpose ?
“It’s hard to fathom how this could possibly be Miljoenhuizen.nl’s fault”
It’s Zwartepoorte fault for being incompetent. It’s no one elses fault that Zwartepoorte is incompetent so they shouldn’t try to blame others.
Re: Re:
It also seems that this incompetence is contagious. The court seems to have caught it. I hope I don’t catch this incompetence pandemic that’s going around.
Wow, still using Google?
Wow, are you losers still using Google? That’s old news dudes. Haven’t you heard, Bing is the new Google because it has a background image on it’s homepage?
Re: Wow, still using Google?
lame
Yet another incompetent judge
Oh look, you found another one! Yet another case of a judge who shouldn’t have taken a case he knew nothing about.
Hm
Did the client first simply *ask* for them to make the change? If so, was the answer something like “we can’t do it” or “we won’t do it”?
If so, a lawsuit was correct next step. And they sued the correct party. And the finding of the court was that “yes, it could be done” with a judgment of “do it”.
Not enough info to know. Would seem to me to hinge on whether the client asked for the changes and what the website’s response was.
I find it difficult to believe that the first reaction of the client was to sue. It’s expensive and, more importantly, time consuming.
I would bet there was a back and forth of emails first. With growing frustration on the part of the client that this was not being fixed.
Especially since the solution would have taken all of a minute to implement. And the only people in a position to implement it *is* the website.
This seems like it *could* be a perfectly reasonable case, but the details will tell.
Can anyone read the court case for actual details, rather than simply re-quoting already info-scarce articles and blogs? Following the links in the articles did not result in high quality data.