How to Unravel the Fake News Networks

A lot of people I have talked to about this project feel overwhelmed by the scale of the fake news problem. There are vast numbers of sites out there at all levels of the fake news spectrum. Whole networks of sites repeat and build on each others’ stories, and professional fake news creators set up brand new ones at will.

According to a survey by Pew research, a majority of people feel a great deal of confusion over basic facts and are not sure whom to trust. In fact, 88% of people reported feeling at least some confusion. It’s only natural to wonder how you can prove anything in this situation.

Trying to solve the entire problem all at once is like boiling the ocean, but fortunately, like most large problems, this one can be broken down to smaller, more accessible pieces.

Break it down by story

Barring exceptions like self-described satire or humor sites, it’s usually incredibly difficult to prove or disprove a blanket assertion like “site x publishes fake news”. Such statements are likely to simply reinforce existing beliefs or fall on deaf ears. They don't drive informed debate and are not likely to change any minds.

On the other hand it's much easier to reason about individual articles, or better yet, individual statements with an article. In my experience, a lot of fake news is built on just a few factual errors with a layer of commentary and opinion added on top. One of the most surprising things I’ve discovered is how many stories can be disproven simply by carefully reading the very sources they cite. Errors are often introduced accidentally or intentionally by authors who misunderstand or misquote the source, or take things out of context. I’ll cover examples of these in more detail in a future blog post.

Once the community has found and highlighted errors in an article, any other reader can quickly and easily verify it. For example, it might take a fair amount of insight and research skill to find the original version of a manipulated photo, or track down the full text of a speech quoted out of context, but once pointed out it’s easy for anyone else to see it.

Find the origin, disentangle the web

A lot of news sources and blogs aggregate, republish, or comment on other news sources. This can lead to readers seeing the same information from a number of different sources, even when they all lead to a single article. Worse, like a game of telephone, every step in the chain can add its own inaccuracies. These articles can persist even after the original article has been retracted or corrected. makes it easy to unravel these chains because community members can locate and highlight the original article, and point out the distortions along the way. In the near future, we’ll also be rolling out additional tools to automatically highlight when an article you’re looking at cites a false article as its source, and reveal the whole web of misinformation. 

Transparency, not censorship

It’s all too tempting to insist that the Government or Facebook or Google must block fake news. In fact, Germany recently passed a law that in effect requires social media sites to police their content under penalty of heavy fines. Although well intentioned this can have unintended consqeuences

I believe that an informed audience that refuses to spread fake news will be far more effective in the long run than any attempts to ban it or deputize Facebook or Google to act as gatekeepers. isn’t a black box that labels something “fake”, but a platform for transparency that helps you figure it out on your own.