« Race, Sports, and Hustle | Main | Academics & Search Engines »
October 27, 2006
Political Google-Bombing
As the midterms approach, yet another tiresome tactic is polluting the information ecosystem: the manipulation of search engine results to paint an unflattering picture of political opponents. Generally called "Googlebombing," I think this development is only good for one reason: it will call people's attention to the politics of search engines as all-purpose information finders.
Search engines can suppress controversy, and at first glance the tactics of the "Google Bombers" may seem an admirable effort to bring salience to oft-overlooked stories:
If things go as planned for liberal bloggers in the next few weeks, searching Google for “Jon Kyl,” the Republican senator from Arizona now running for re-election, will produce high among the returns a link to an April 13 article from The Phoenix New Times, an alternative weekly. Mr. Kyl “has spent his time in Washington kowtowing to the Bush administration and the radical right,” the article suggests, “very often to the detriment of Arizonans.”
But such efforts ignore the value of organic search results untainted by the classic forces of power and profit. I don't mind if paid results are determined by who has the most money or best way to manipulate the system, but organic results at a general purpose search engine are held out to be an accurate account of what is the case. A search for Jon Kyl should, ideally, produce among the highly ranked organic results some relatively balanced portrayals that give voters a good sense of what he stands for. It is deeply troubling to think that a cabal of his opponents (or supporters) could crowd such results off the top pages in order to advance their agenda. The manipulators also risk provoking search engines into downranking political sites, or at least those heavily interlinked ones which (look like they) are trying to artificially affect the results.
The groups profiled in the article should also know that they can easily be outmatched in the production of digital astroturf. Yet one of their ringleaders sanguinely "hopes[more] political campaigns would take up the tactic, which he called 'search engine optimization,' as a standard part of their arsenal." I admit that search engine optimization is a very complex topic, but it often boils down to the commodification of salience: if you give enough money to the SEO, they try to get you ranked high in response to certain queries. Given the already overwhelming influence of the "dollar primary," the last thing we need to do is to extend that dynamic into the world of online politics.
PS: The speakers at this panel had some interesting insights on the topic...more on that later. I can already sense some people will think I'm mistaking Google for Wikipedia. But one can spot an unhealthy dynamic without necessarily committing to a particular way of stopping it. I suppose my biggest beef here is with the sense that an "arms race" of google-bombing is inevitable/desirable.
Posted by Frank Pasquale at October 27, 2006 09:16 PM
Trackback Pings
TrackBack URL for this entry:
http://www.concurringopinions.com/movabletype/mt-tb.cgi/1394.
Comments
hmm. Some of the reasoning comes dangerously close to suggesting that search engines are an essential facility. Can search engines have a monopoly on information? Should they be regulated?
What's the alternative? Barrier of entry is low, create your own that does a better job.
Posted by: Jack S. at October 28, 2006 12:23 AM
I don't think any particular search engine is an essential facility--yet. But I do think that the cultural/political importance of dominant search engines' results is such that we ought to be reluctant to let them be decided in a black box manner, completely immune to scrutiny of any kind.
As for "start your own search engine;" well, there are a number of barriers to entry:
1) copyright in content: the dominant search engines may have negotiated licenses for content, which you can afford.
2) copyrighted/patented/trade secret-ed algorithms: you may need to license search software from the dominant SE's, who may not be inclined to let you.
3) consumer habit: people are used to using a certain number of providers, and are pretty loathe to switch, despite pretty interesting alternatives like Clusty and Ask.com. Google did manage to displace Yahoo!, but only after developing much better technology. If we've hit some sort of "flat of the curve" of search innovation, there may not be much of a chance for a new entrant to knock off incumbents.
4) "rich get richer" dynamic: the more searches an engine gets, the more likely it can use studies of extant behavior in response to results to sharpen its algorithm.
Posted by: Frank at October 28, 2006 03:42 PM
If political Google-bombing gets out of control, what are the odds that Google will tweak its algorithms to reduce the effects of such behavior?
Meanwhile, you say "A search for Jon Kyl should, ideally, produce among the highly ranked organic results some relatively balanced portrayals that give voters a good sense of what he stands for." Why is this the case? Do you have empirical data about, special insight into, what a searcher for Jon Kyl is looking for?
Eric.
Posted by: Eric Goldman at October 28, 2006 06:12 PM
Eric,
1. Yes, I think Google will tweak its algorithm to control the situation. But I think its method of doing so is a matter of public concern. Will it simply give less weight to densely interlinked political sites? Will it give less weight to all political sites? How might "political site" be defined for these purposes?
I'm not saying all this has to be made public; that would just help the spammers/bombers more. I'm just saying that perhaps there should be some input into this process to assure it's reasonably responsive to the public interest. Historically, we have been concerned (and regulated) when a given medium (like broadcast TV) turned stably oligopolistic.
2. As for my insight on "what a searcher for Jon Kyl is looking for"--no, I can't read minds! But we can work backwards to develop some sense of the legitimacy of balance in results.
a) If my query were "Jon Kyl scandals," then I'd probably want the New Times article mentioned above.
b) If my query were "Jon Kyl achievements and honors," I'd probably want more positive treatments.
c) "Jon Kyl" standing alone probably indicates a searcher trying to get some sense of what are the best resources out there on Jon Kyl.
My bottom line, though, is that some SE results ought not reflect only what a searcher wants, but also what that searcher ought to want. A consumer-satisfaction metric of quality is fine for searches for toaster ovens, but does not capture the full range of values that ought to infuse democratic dialogue. Just as I don't think we can deem a campaign full of substanceless attack ads legitimate even if they satisfy the preferences of a vast majority of voters, I don't think high-ranking SE results that are wholly partisan or manipulated or sensationalistic realize the promise of this new medium.
PS: There is, of course, the issue of "many Jon Kyls". I think a site like this suggests a potential response to the problem:
http://www.froomkin.com/
Though I know that directory-based SE's are largely a thing of the past (but see dmoz.org!).
Posted by: Frank at October 28, 2006 06:40 PM
