First publicized in a Wired article, WikiScanner project aims to reveal conflicts of interest in Wikipedia editing. The Wikipedia model for putting together an encyclopedia has already been criticized endlessly on two grounds. First that a “coalition of aspiring experts” could not put together a reference work to compete with the existing stalwarts of knowledge. Second, complete openness to input from anyone and the reliance on self-policing to correct abuse means that at any point the loudest, most persistent voice is more likely to be hear than the most informed.
Wikiscanner does not answer the question of how vigilante the band of volunteers have been at upholding the so-called neutral point of view. But it can provide an interesting perspective on the extent neutrality has been under siege. Since Wikipedia keeps a history of all edits, the task boils down to tracking the contributors against the entries they modified. Not surprisingly the project found that users from Diebold, Wal-Mart, Dell and various congressional offices all edited entries related to their employer, predictably removing critical passages or providing counter-points in defense.
This is a good start but fails to answer two important questions:
- In each case, were these instances of over-zealous employees (as Dell seems to have argued in response to a New York Times article covering the same problem) or a systematic campaign endorsed by the company itself– or both?
- Is there any reason to believe that all conflicts of interest can be identified? WikiScanner depends on one piece of information: IP address of the contributor. At this point, natural selection favors those who realized that IP addresses can be tracked back to their source organization using a simple whois query. Better organized attempts to subvert Wikipedia content could originate from home IPs or anonymizing proxies. In that sense WikiScanner is another instance of the state of cyberspace accountability: “we catch the dumb ones.”