Monday, July 10, 2006

Is Wikipedia criticism justified?

(Via Slashdot) Frank Ahrens of the Washington Post adds to the growing criticism of Wikipedia for its lack of accuracy at all times. The case in point this time around is the death of Kenneth Lay, and the ensuing sequence of often erroneous updates to Wikipedia's Lay entry. The chronicle of events described by Ahrens paints a very interesting picture:

10am, July 5, 2006: News organizations report Lay's death due to an apparent heart attack.
10:06am: First Wikipedia update on Lay. Wrongly claims his death was "an apparent suicide".
10:08am: Updated to say cause of death was "an apparent heart attack or suicide".
10:08am: Updated again to say cause of death is "yet to be determined".
10:11am: Comment added to imply that guilt from the Enron scandal caused the suicide.
10:12am: Cause of death correctly identified as being due to massive coronary, as reported by Lay's pastor.
10:39am: More speculation as to the cause of the heart attack, but clearly identified as such.
Afternoon: Stable entry with correct facts.

Ahrens looks upon this series of events as a negative -- proof of the inherent untrustworthiness of Wikipedia -- and even goes so far as to call it "an active deception, a powerful piece of agitprop...".

This attitute would appear to get the whole story backwards for two reasons:
  • First, the story here is not that anyone can deface Wikipedia with blatant untruths. (After all, that is an obvious consequence of global editability.) The story is the rapidity with which correct information percolates into the system and falsehood is eliminated! It amazes me that the information on Lay was updated within twelve minutes of his death and that it was fact-checked and corrected within a few hours. What other knowledge repository has this kind of latency to correct information?

  • Second, Wikipedia should be treated as what it truly is, not as what we'd like it to be in an idealized universe. What Wikipedia is, is an unimaginably broad repository of community-edited documents that is generally accurate on most matters. What it is not is the final word on a subject that bears the reputation of well-known editors or publishers behind it. Sure it would be nice if Wikipedia could also have the latter characteristics but it does not; if a reader assumes that it does, the fault likes with the reader, not with Wikipedia.
I do see interesting efforts under way to lend Wikipedia more of the reliability associated with editorial oversight and make it even more useful than it is today. Articles within Wikipedia undergo different levels of peer and editorial reviews and make their way into an upper echelon "core" that is tagged as such.

It would certainly help if more of an effort was made to clearly label each article with the level of peer review it has undergone, and the amount of stability that it has achieved. One can even conceive of simple automated evaluations based on the recency and frequency of updates to provide the reader with better guidance on the likely accuracy of an article.