Wikipedia FINALLY Exposed. The Wiki Bias is Undeniable.

[00:00:00]Let's let's let's just start at the very beginning. How did you even get into this? You're an investigative journalist. What made you decide to start investigating Wikipedia?

[00:00:09]>> It was really kind of a coincidence. I was working with someone um a really great researcher, a machine learning researcher named David Rosado who did this very kind of simple. The results were simple. It was a study about Wikipedia bias. And the results showed that there's a clear bias towards the left. So if you take like um American politicians, Supreme Court justices, even journalists that are mentioned on Wikipedia, they tend to be associated with more favorable language on the platform. And on the right, you have got the opposite effect. They tend to be people on the right tend to be associated with negative terms and and negative sentiment on Wikipedia. I was like, okay, we all kind of know that you kind of like have a gut sense that that's the case, but why? That's what I wanted to understand. How did it get to that point? How does it work and what does it mean for the bigger information ecosystem?

[00:01:02]>> Okay, let's let's start first with some examples of what you're talking about.

[00:01:08]Uh, and again, you know, I could show mine, but that's just so nominal. I mean, we're talking about >> Trump, Elon Musk, Iran, Israel. I I mean, this is literally impacting the way the globe thinks about things.

[00:01:24]Wikipedia is not just available in America. So, can you give the audience some glaring examples of what you're talking about that they might be able to identify and recognize immediately?

[00:01:35]>> Yeah, I mean there's so many. It's the entire platform in a way. And just to your last point that you kind of touched on, it is the globe. It's the entire information ecosystem. It's what feeds Google results. And most important of all, it's what feeds AI disproportionately, more so than any other single knowledge platform. The AIs that we are relying on today get their training from the training data from Wikipedia and they get their reference data when you ask a question. Oftent times it will pull the data straight from Wikipedia. So on Trump for example, um I did some reporting about how a small cluster of editors on the platform worked to cast Trump as an authoritarian, which is a very extreme claim to make about an American president, even when you may really really dislike. To call him an authoritarian in an encyclopedia, we're not expressing an opinion on X or Instagram. Um, these editors relied on Marxist sourcing, far-left hard-left sources, sources that were working with Obama, former Obama officials in NOS's.

[00:02:44]They launder it through Wikipedia, but you, as the user of the site, don't really know where the information is coming from. you just get this blank statement and then you go over to like the supreme leader of Iran at the time at least or um Xiinping the supreme the the uh whatever they call him the paramount leader the the leader of China and you don't see the term authoritarian they call Trump an authoritarian but not an actual authoritarian this is just one example in the Iranian stuff we have systematic infiltration by Iranian media state media IRGC these as our terror supporters. Um, the biggest sponsor, state sponsor of terrorism in the world.

[00:03:25]Wikipedia relies on their data, on their information to prove facts about what you, as a consumer in the United States, are learning about the world. We see the same thing with China. We see the same thing even with al Qaeda media being cited as sources on Wikipedia. It goes on and on and on. The more you dig, the more you find. This used to be like kind of a joke. Oh, your Wikipedia page a decade ago. Um, and then I'll say this, I started to notice it taking a very bizarre turn when it came to my reputation and my business partner and I would struggle to try to clean up the page and we would find it impossible and it had been like completely commandeered. But the part where this is relevant and not just my personal sour grapes about this is at the time I was working with the Clinton Foundation and literally working directly with the Clintons in the Clinton Global Initiative with their health matters program and their people were like, "Oh yeah, no problem. We can lock the page."

[00:04:33]>> And they locked my page >> for a long time by the way so that nobody could mess with it. and I never thought anything of it >> until now. So, let's let's then table that and go back to the beginning. How did this start? Was this intended to be a propaganda machine utilized by bad actors?

[00:04:57]How does this happen? How did we get here?

[00:05:01]>> Great question. And we're we're definitely going to have to come back to the Clinton piece because it's really really significant. But to answer the question of how we got here and where it came from, Wikipedia started in the early 2000s. It was 2001 and it was the idea was kind of to put an encyclopedia on the internet which is a very novel idea at the time. Um they did that initially by trying to do it in a very traditional manner. They had editors and review process and they found out it was extremely slow and extremely expensive to do it that way. And then I believe it was Larry Sanger had the idea was a co-founder of Wikipedia had the idea of using this concept of a wiki which was a website that anybody could edit and marrying it with the concept of a encyclopedia so that anybody could edit the encyclopedia and it just it took off. It was lightning in a bottle. And because of this incredible growth rate that they saw early days and that sustained itself for quite a long time, it really became this amazing media story. Like for the media, it was novelty. It was internet. Back in the day, today the media kind of has like an this kind of uh animus against tech, like kind of an anti-tech attitude that pivoted after the first Trump election.

[00:06:17]But before then it was very the the media was largely pro tech and anything that was seen as a tech innovation was really celebrated and they hyped it up and they made it out to be this like great solution to the age-old question of how do we know what we know? Um and it seemed like we had a solution. If you crowdsource the knowledge and you get a bunch of different perspectives, you can actually flatten out the bias and the noise and elevate the signal. like in on paper it's a pretty good idea and sort of it worked. I mean already early early days you had problems. There was a famous case with a journalist named John Seanthaylor who was a very senior journalist in in in American in the American press. And someone just decided to to take to his Wikipedia page in the early 2000s and lie and say he was uh involved in the assassination of Bobby and uh John F.

[00:07:14]Kennedy. And he happened to actually be a close friend of Bobby Kennedy. So this was a a slander that hurt him personally and he was one of the first to really raise the red flag to say something is wrong with this because he couldn't get the page changed. He was as important and distinguished and influential figure in American society as you could possibly be and he had no recourse and that's what really outraged him. So over the years the story built and built and built and we had this really great myth about Wikipedia.

[00:07:48]Somewhere along the lines 2015, 2016, we started to understand that the internet is not just fun and games anymore. And knowledge platforms like Wikipedia were susceptible to all kinds of manipulation, to hijacking, to ideological manipulation, to statebacked or state- linked operations. Um, and this only increased as Wikipedia positioned itself as what it described as part of global knowledge infrastructure. That's something that happened around 2016. A very deliberate pivot. That is where we have the Clinton um element in the story.

[00:08:28]>> I see. Okay. All right, guys. If you like this clip and you want to watch the full episode, click here. And if you want more content like this, you can subscribe to the show by clicking