Truthiness showdown: Google's "Knol" vs. Wikipedia

Google is planning a new system to distribute human knowledge. Will it kill Wikipedia?

By Farhad Manjoo

Published December 14, 2007 7:27PM (EST)

Having just written a book about how digital technology is changing cultural ideas about truth -- shameless plug: to be released mid-March from Wiley; pre-order here -- I'm fascinated by Google's announcement, late yesterday, of a Wikipedia-like application called Knol.

Knol's goal, writes Udi Manber, Google's engineering chief, in a blog post, is "to encourage people who know a particular subject to write an authoritative article about it." The system, which is currently running in an invitation-only beta, offers free Web hosting space and editing tools to allow anyone to write up a page about whatever they like. Google is calling each article a "knol," which it says stands for a "unit of knowledge."

Each knol will be something like a one-entry blog on a single subject. Google hopes folks will create many of them, building a corpus of human knowledge.

"The goal is for knols to cover all topics, from scientific concepts, to medical information, from geographical and historical, to entertainment, from product information, to how-to-fix-it instructions," Manber writes. "Google will not serve as an editor in any way, and will not bless any content. All editorial responsibilities and control will rest with the authors."

Some of this may sound familiar. Search the Web today for what Google grandly calls "knowledge" -- say, for human growth hormone or Grover Cleveland or Boney M or L. Ron Hubbard -- and up near the top you'll see a link to a site you've likely dropped in on a couple times or so. Right, Wikipedia.

So the question of the day: Is Google going after Wikipedia and the many Wikipedia-like sites to come up recently -- Squidoo, say, or Mahalo?

Surely Wikipedia is in Google's sights; that site gets huge traffic, and Google wouldn't balk at traffic, especially as it will display ads on knol pages (with each author's permission; authors would share in the revenue generated by the ads).

But Knol differs in a key way from the popular everyone-pitches-in encyclopedia, and as a result, both could prosper, offering different bits of knowledge for different purposes, and different audiences. In fact, by adding more experts to the Web, Knol, if it takes off, will probably improve Wikipedia, not kill it off.

Wikipedia relies on "collective" knowledge. Each Wikipedia article is not the creation of one author but multiple authors who are, though not technically anonymous, not given explicit credit for their contributions. The Wikipedia article on human growth hormone, for instance, offers -- or, attempts to offer -- the Web's combined information on the subject. No single expert's ideas are given prominence; in fact, among Wikipedia's founding tenets are "neutral point of view" and "no original research," principles meant to diminish strong personal opinions.

But "the key idea behind the knol project is to highlight authors," Manber writes. So you wouldn't find just one knol entry on human growth hormone that attempts to represent the collective view. There would be several knols on HGH, and each one would feature the author's name prominently -- and each would offer that author's opinions, analysis, judgments and interpretations on the subject.

Manber says that each knol would be subject to "community" input. "People will be able to submit comments, questions, edits, additional content, and so on," he writes. "Anyone will be able to rate a knol or write a review of it."

These systems will give you some way to judge the knol, but your basic approach to the article -- your decision over whether to trust what it says -- will be very different from your approach to an article you find on Wikipedia.

Wikipedia's veracity rests on verifiability -- you determine the truth of any particular fact you find there by checking the citations that back it up. And even if you don't take the time to check the citations, you can assume that a fact's continued presence on the site means it's true, because others have probably checked it (although, of course, caveat emptor).

A knol, on the other hand, is based on individual expertise: You trust what's in the knol not only because the author provides citations for everything she says, but also because, more fundamentally, you're convinced that the author knows what she's talking about. To help convince you, the author may list her credentials, may explain her experience, may tell you about experiments she has personally conducted. If those things sit well with you, you cast your lot with her.

To understand the likely differences between knols and Wikipedia pages, take a look at the screen shot of the sample knol that Google put up. The page -- created by Rachel Manber, director of the Insomnia & Behavioral Sleep Medicine Institute at Stanford University School of Medicine (presumably Rachem Manber is some relation to Udi) -- is titled "Insomnia: Risk factors, diagnosis, and treatment."

It offers a thorough, expert's guide to the many treatments you might try to alleviate your sleeping disorder. The article is particularly helpful in explaining the efficacy of drugs compared to that of behavioral therapy -- it explains that drugs work faster, but that therapy lasts longer.

The information here is highly trustworthy -- after all, it's written by a Stanford expert on insomnia, and, for further guidance, you can check out one of many peer reviews of the piece.

The Wikipedia article on insomnia, by contrast, offers a much less detailed picture of the sleep disorder. Mainly, it glosses over the different kinds of insomnia, and lists the causes and treatments. It will not be helpful for people looking to get a better night's sleep.

On the other hand, the Wikipedia page does offer a wider field of view -- it collects both expert and non-expert information, and thus, for instance, mentions that "cannabis has also been suggested as a very effective treatment for insomnia."

For proof it cites the work of the late Tod Mikuriya, a psychiatrist and leading proponent of marijuana for medical purposes. If he were alive, Mikuriya could have launched his own knol pushing cannabis as a cure for insomnia, and, if you were convinced, you could decide to go with his treatments rather than Manber's.

But here's the key thing: If competing knols do indeed proliferate, Wikipedia would grow to include the knowledge they contain -- and, in that way, it would only get stronger. The more stuff that goes online, the bigger Wikipedia grows.

So Google isn't killing Wikipedia. It's helping it.

Farhad Manjoo

Farhad Manjoo is a Salon staff writer and the author of True Enough: Learning to Live in a Post-Fact Society.

MORE FROM Farhad Manjoo

Related Topics ------------------------------------------