Tuesday, February 9, 2010

Media: Wikitruth

TORONTO, ONTARIO - I've written before on this blog about my frustration with people that take information on the Internet more seriously than actual primary sources. It turns out that the root source of this trend had already been written about extensively and I had just never bothered to read it. Well-known Boston technology writer Simson L. Garfinkel wrote an article in the MIT Technology Review (unfortunately not available for free) that's the best I've seen describing the Wikipedia standards for truth and the problems they create.

Wikipedia's standards for inclusion basically come down to verifiability. The best source of information in the Wikipedia world is a web page. The best web page is in the English language. So, if there's an English-language web page with the information, it can be accepted as a reference, even if the content is completely wrong.

The really disturbing thing is that people finding mistakes in the Wikipedia world cannot actually just fix them, as Wikipedia touts. If the article is on something that you have any sort of connection to--most profoundly an article on you, but also any organization you might belong to or anything you might be perceived as having a financial stake in--then you are considered a biased source, and you are not allowed to fix it and your English-language web page cannot be considered valid source material. This might make sense for subjective aspects such as whether a public incident constitutes a scandal, but it also means you can't correct your own birthdate, the number of children you have, or your profession. Is there really any bias possible in things that are just incontrovertible facts?

So, basically what the Wikipedia view world does is reinforce that whoever is most vocal in the media space speaks the truth. If one can get one's friends to post an apparently-unrelated-to-you assessment of your life that is actually completely fabricated, that would be accepted in Wikipedia as a valid source of information. But, if you can't convince a third-party to post on the web that information on someone else's web site about you is slander, then that slander is considered the truth, not the actual reality.

This personal aspect may not be of much concern to the average person, but the problem is much broader. If a web site decides to publish that the MD-80 was an Airbus aircraft, then Wikipedia can repeat that as truth. That might seem absurd, since it is common knowledge that it was a McDonnell-Douglas aircraft derived from the Douglas DC-9, and in this case those correcting the article (as long as they aren't associated with successor Boeing!) could cite other web sites with the correct information and the fix would be accepted. But, on more obscure topics, on which information is not readily found on the web, the information would become accepted truth. This is an especially large potential problem in science, as much scientific data is currently available on-line only through paid peer-reviewed journal sites and therefore isn't valid as Wikipedia references. As far as the Wikipedia world is concerned, the best peer-reviewed version of scientific truth doesn't exist.

It's not hard to understand where the phenomena I take issue with comes from in this world. People that turn to Wikipedia for information implicitly accept the Wikipedia version of truth, even if they would disagree with the policy. Generation Y and younger, growing up in an open-source and crowd-sourced world, have minds that work the same way that Wikipedia does, in which if they can find the information, then that's the information they will accept. There's no sense in further investigation to figure out the validity of the information--who has time for that?

The implication of this conception of truth in a world in which governments (not just those as aggressive as China) and commercial interests (the whole open access controversy is about commercial interests favoring certain web sources over others) increasingly control the Internet is profound. It may actually be harder to find accurate facts on the future Internet than it was before we had that resource and had to turn to things like researched encyclopedias.

One obvious action that will help is working to make scientific information freely available as quickly as possible so that it can be cited outside of research communities--that should be easy to solve, but the broader problem is not so easy. Garfinkel suggests not relying on crowd-sourced information such as Wikipedia, period. Yet, one wonders if there actually be any choice. If Wikipedia doesn't modify aspects of its policies on what constitutes valid information, it may have serious long-term negative consequences for society. I think the consequences are already appearing in the guise of people accepting the Wikipedia version of truth as their own world view.

No comments: