18 Comments
User's avatar
Dave92f1's avatar

How could it be otherwise? Wikipedia is written and edited by volunteers, and naturally reflects the views of those volunteers.

Left of center people are generally far more literate than right of center people - more have college degrees, more of those are in journalism or literature or politics. (Exactly as with the media in general.)

None of that makes them smarter or more likely to be correct, but it does make them far more likely to edit Wikipedia (and to be in the media industry).

Perhaps right-of-center people are more mathematical (this is pure speculation) or practical, or honest, or ...whatever. But none of that makes them likely to spend time writing articles for Wikipedia.

Expand full comment
TonyZa's avatar

True. There are some forks of Wikipedia that try to be more objective or right wing but they never took off.

My take is that right-of-center people are more likely to be religious so they focus their charity and volunteer activity in churches while leftists engage more in political activism.

Expand full comment
Taylor's avatar

“Literate” is not the right word.

Expand full comment
RDM's avatar

What a massive, gaping hole this article fills so superbly!

Expand full comment
James Tucker's avatar

Hilarious the the ADL's hatelist is reliable, but they are banned as a source on Israel-Palestine. It would be hard to find a better example of blatant leftist bias than that.

Expand full comment
two wheeler's avatar

"Based on this take, the key information battle to come is that of who gets to decide what bias goes into the LLMs"

interesting study but unnecessary: Larry Sanger, co-founder of Wikipedia made the same point not so long ago, saying the site only gives "an establishment point of view", which was not what he had intended when he started it. He explained, "If only one version of the facts is allowed, then that gives a huge incentive to wealthy and powerful people to seize control of things like Wikipedia in order to shore up their power." Ditto for AI.

Expand full comment
Emil O. W. Kirkegaard's avatar

Quantitative analysis is king, Sanger was just talking, Rozado did the work.

Expand full comment
Mike Hind's avatar

In my own most recent piece I contrasted the Wikipedia opening sentences for 'extreme right' and 'extreme left'. You can guess already which one was freighted with fear and which was neutral.

Expand full comment
Anatoly Karlin's avatar

As they say, reality has a liberal bias. The arc of history bends towards liberalism.

There is no way to "positively" cover Republican politicians and the misogyny, Christian nationalism, and pseudoscience they stand for while maintaining moral and intellectual consistency.

Expand full comment
r1based's avatar

You could say the same thing about leftist hypocrisy: a left-wing topic couldn't be covered positively given all the hypocrisy, while maintaining intellectual consistency. But you can, and they do, by just glossing over it. Anything can be covered positively with enough imagination

Expand full comment
Chriss's avatar

I detect some bias in this comment.

Expand full comment
Taylor's avatar

Reality has an *anti*-liberal bias.

Example: Wearing a skirt and calling yourself a woman doesn’t make you one.

Another one: Race is real and detectable. Everyone does this, especially liberals when targeting White people.

Another one: There are innate, biological differences in behavior between men and women.

And so on. The list is endless.

Expand full comment
Eugine Nier's avatar

Quick sanity check: How many genders are there?

Expand full comment
Anatoly Karlin's avatar

Arbitrarily many.

Expand full comment
Taylor's avatar

See? You’re insane.

Expand full comment
Realist's avatar

I see no surprises.

Thanks for the confirmation.

Expand full comment
Mark's avatar

In the German version I was surprised/shocked reading "Schlüsselreiz" - esp. when it came to the topic of 'super-stimulus' . While the English version states very reasonably: "Lorenz and Tinbergen accounted for the supernormal stimulus effect in terms of the concept of the innate releasing mechanism; however this concept is no longer widely used.[citation needed] The core observation that simple features of stimuli may be sufficient to trigger a complex response remains valid, however." - The German version make it sound much less "valid", sadly.

https://de.wikipedia.org/wiki/Schl%C3%BCsselreiz#Kontroverse

Btw: Konrad Lorenz and his golden boy, I. Eibl-Eibesfeldt, were no wokies. Lorenz became a member of the NSDAP, and Eibl wrote in 1998: „ ... eine Begrenzung der Immigration aus kulturell und anthropologisch ferner stehenden Populationen notwendig." The "Left" never liked them, Skinner was much more to their bland/k slate taste.

Expand full comment
User's avatar
Comment deleted
Jun 24
Comment deleted
Expand full comment
Emil O. W. Kirkegaard's avatar

I proposed something like this a while ago. Apparently, this was before AI/LLMs, so I wasn't farsighted enough. Maybe it's time to re-iterate the idea. As you already wrote the post, I suppose you should post it. :)

https://emilkirkegaard.dk/en/2018/04/wanted-scientific-immune-system-to-identify-weak-studies-getting-lots-of-attention/

Expand full comment