> [W]hen people thought the Earth was flat, they were wrong. When people thought the Earth was spherical, they were wrong. But if you think that thinking the Earth is spherical is just as wrong as thinking the Earth is flat, then your view is wronger than both of them put together.
Everything is biased. Everyone, every single human being, is aligned.
Of course but as discussed many times here before, Wikipedia leans left - presumably reflecting the statistical properties of the demographic of the people drawn to edit and moderate it - as implied by the comment you are replying to - and that can be a significant issue for topics (e.g cultural, historical, social, political etc) where that bias filters what might be assumed by the user to be objective answers.
This isn't a left vs right thing either; there are plenty of publications, demographics and institutions that lean right. The problem is the transparency, awareness and communication of that bias when using them as sources for tools like this.
In the underlying study, there is no mention of the word "bias".
Here's a sample quote which is also concerning:
For recent topics, we look at the most edited
Wikipedia articles 1
in the first four months of 2023.
The number of edits is also a good proxy for how
much interest there is around a certain topic.
True - and it may also be an indication of a topic that is heavily contested. If the two (or more) views on the "truth" of the article are imbalanced, the chatbot will reflect that imbalance, and can therefore in no way be said to “outperform all baselines on factual accuracy".
To be fair to the researchers, they do address related concerns and talk about avoiding some areas of discussion, but the headline here is extremely misleading.
That is subjective, and depends on where you think the "centre" is.
I don't regard Wikipedia as reliable on any topic that is political or involves national history. Modern Wikipedia expects editors to support their edits with citations to "reliable sources", which means the mainstream press, mainly (because primary sources are deprecated). But the mainstream press is overwhelmingly right-wing, and left-wing papers and magazines are usually explicitly rejected as not reliable.
On matters of politics and history, I always dig into the citations (unless I'm happy to get a sketchy version that isn't really accurate). But on most technical and humanities-based topics, the articles are usually quite good (and often much deeper than 1").
There's still way too much stuff in articles that is not cited at all. That changes gradually, as editors delete uncited material, and others come along with suitable citations. I think it's getting better all the time.
That is subjective, and depends on where you think the "centre" is.
Not at all. Even wikipedia itself acknowledges it [1] - and you can bet the editors responsible for the bias were fighting tooth and nail against that admission - which gives some idea how unbalanced it must be in reality.
Modern Wikipedia expects editors to support their edits with citations to "reliable sources", which means the mainstream press...
And academia - don't forget academia, that bastion of the right.
...the mainstream press is overwhelmingly right-wing
That's ridiculous - The Guardian?? The Washington Post? New York Times?
I think you've made a point about Wikipedia though, but perhaps not the one you intended...
> That's ridiculous - The Guardian?? The Washington Post? New York Times?
Yes, those. I'm not closely familiar with NYT and Washpo; but I'm a long-time reader of the Guardian. It's role is to delimit the left end of the Overton Window, which is roughly the left-ish end of the Shadow Cabinet, and is currently well to the right of Labour Party membership. From my limited exposure, I'd say NYT and Washpo are well to the right of The Guardian, as the Democratic Party is well to the right of Labour in the UK.
Your WP link doesn't speak to any strong bias in English Wikipedia; it says roughly that WP content is "establishment", which isn't surprising, considering the Reliable Sources policy. It also says that older articles exhibit a stronger "left-wing" bias (i.e. pro-Democratic). And the article is written from a US POV; in the USA, socialism is still a "bad word", and the word "liberal" is used instead.
If you consider the likes of The Guardian right-wing there is really no point in continuing this discussion as I don't believe it can be done in a good faith manner.
The Guardian has been a supporter of the Israeli government since before there was an Israeli government. Nobody could accuse the Israeli government of being left-wing.
The Guardian was part of the coalition that hounded Jeremy Corbyn out of the Labour leadership, because he was too "left wing". Starmer refused to let him stand in the seat he's represented for decades; The Guardian strongly backs Starmer, who represents the most right-wing elements in Labour.
I'm happy to discontinue this discussion, not because I suspect you of bad faith, but because political discussions aren't generally on-topic on HN.
That said, bias towards, and alignment with, verifiable reality is possible to achieve, and getting there partway is better than not at all:
https://hermiene.net/essays-trans/relativity_of_wrong.html
> [W]hen people thought the Earth was flat, they were wrong. When people thought the Earth was spherical, they were wrong. But if you think that thinking the Earth is spherical is just as wrong as thinking the Earth is flat, then your view is wronger than both of them put together.