Why Filtering Fake News on Facebook Is a Slippery Slope

I remember seeing an article on Facebook about some catastrophic news regarding Hillary’s emails. I was shocked and quickly clicked the link to read the article. When I arrived on the website, the content was so outlandish I began to think I was being hoaxed. I looked for a second source providing the same information and found none. Indeed, I had been reading “fake news.”

It’s been suggested that fake news — such as a story about the Pope endorsing Donald Trump — tilted the results of the election.

Mark Zuckerberg denied such claims, saying that more than 99% of what people see is authentic and that only a very small amount is fake news.

However his remarks haven’t stopped a small team of Facebook employees forming a task force with the purpose of questioning Facebook’s role in promoting fake news.

Censoring fake news sounds like a great idea. After all, it’s fake. Who wouldn’t want to get rid of fake news? But the issue is far more complex. It calls into question our ability to communicate important information about the most pressing issues facing our society today. This article covers the potential downsides of censorship on social media and suggests alternatives.

How do we determine what’s fake?

During the election, news that was said to be fake turned out to be true, and news that was trusted as true turned out to be false.

Concerns over Hillary Clinton’s health were called “conspiracy theories.” That was until there was insurmountable evidence: her collapse during a September 11th memorial (which was recorded by a citizen journalist). Her team then reveal that she had pneumonia.

Perhaps more alarmingly than “fake news” that turned out to be true, was news that was trusted to be “real,” turned out to be fake. Many pundits said Trump had no chance of winning. They took statements out of context. Regardless of political beliefs, most people seem to agree that the media spent too much time talking about unimportant issues and not enough time on important issues that people cared most about. The fear mongering may have caused more problems than the fake news.

My point here is that it’s very difficult for humans to determine what’s true and what’s not.

Historically, humans have not been very good at determining what’s true.

At the core, humans want to stay alive and reproduce. We don’t necessarily need to know what’s true in order to do that. We can all have different stories in our head and still stay alive and reproduce.

Even if I believe a religion that turns out to be untrue, I will probably be able to stay alive and reproduce. It might even improve my chances because of the community that religions often provide.

Scott Adams provided a great explanation of this challenge in a recent blog post

“…consider that half of Americans believes their country just elected a racist, homophobic, sexist and the other half thinks we elected an open-minded guy who is no more sexist than most people. No matter which half of the country is right, the other half of the country is in a deep hallucination. We just don’t know which half. (Yes, I know that you know the other half is hallucinating. But keep in mind that the other half thinks you are the one hallucinating.)”

Presumably the filtering of fake news would be done by some combination of humans and technology. However cognitive biases may prevent it from being done effective.

The risk in filtering

Before social media, ideas were transmitted primarily through big media. It was much harder for important messages to spread. For example, slavery was at one time a social norm in the United States.

Social media offered the opportunity for anyone with an idea to be heard and for anyone willing, to listen. Now, one-quarter of the world’s population uses Facebook. Hundreds of millions of links are posted to the site every day.

Social media is the communication medium of today. It’s allows good ideas, like opposing slavery, to spread even faster. Given the risks, I’d hate to limit our ability to communicate new ideas and for them to be heard by the masses without an important reason.

What happens when new people take over the filtering?

The size and scope of government has generally been increasing since our country’s founding. Many people are happy about that, particularly when the government is making decisions that they agree with. But when someone gets into power (such as Donald Trump) that they don’t agree with, they’re not so happy.

There’s a risk of giving a company, a small team, or any individual power over speech — something so critical to the human experience. What happens if those people are replaced by people with dangerous political beliefs or prejudice towards certain groups of people? What happens if an event in the life of the person in power causes them to become biased against certain ideas? What if, like most people, they simply aren’t effective at determining what’s true and what’s fake?

Alternative Solutions

The simple but hard solution is for everyone to be more vigilant about judgeing the quality of news. The easy but complex solution is to censor. Per the “explained by marketing” model of the world, the latter is more likely to be chosen because it’s easier. Here are a few alternatives to centralized censorship.

More Social Networks

With more options for social networking, if Facebook starts censoring too much, or incorrectly, and people don’t like that, they could choose another social network that does a better job. Therefore, Facebook would be incentivized to get filtering in line with what users want.

The challenge is that Facebook/Twitter arguably have monopoly power. This is due to network effects. Network effects mean that as more people are on the platform the harder it is to leave or use an alternative. For publishers, consumers are already on Facebook. For consumers, publishers are already on Facebook.

It’s very hard for a new social network to gain critical mass. Therefore it’s hard for a publisher to justify allocating resources there and it’s hard for consumers to justify spending time there until the publishers they consume on Facebook and Twitter are all actively there.

Competition Algorithms

An alternative to creating entirely new platforms is to create new algorithms within existing platforms.

With competition, people could choose which algorithm they like best. When there is competition, there is incentive for people to create better solutions. The downside is that you’re still on Facebook.

A Decentralized Network

I’ve written about Gab.ai previously. Since that post, the network has only grown larger. I trust the team there to maintain bipartisanship and their clearly defined rules about what content is acceptable and what’s not. None the less, it suffers from the same risk as Twitter and Facebook: centralized control.

An open source and/or blockchain based social network without centralized control might be the end game of recent censorship on the various social network. Of course here you run the risk of horrible offensive and/or dangerous content running wild. But it may be better than the alternative of important and peaceful information being censored.


Perhaps we don’t need to be so catastrophic. Perhaps with the right tools, self-censorship on existing platforms can be more effective. Tools like a “Yelp for news,” a browser plugin that blocks fake news sites, or simply more focused feedback mechanisms of Facebook could help.


There always has been and probably always will be “fake news.” Previously it was rumors passed through the grapevine. Now it’s clickbait articles shared on Facebook.

I don’t want people getting false information just as much as the next guy. However, I don’t know that fake news is a problem big enough to justify giving filtering power to a small group of people, or a problem that couldn’t be solved more effectively through other means.

The ability to speak and learn is fundamental to the human experience. Those activities are now primarily done through social media. If recent actions by big social networks towards censorship get more progressive, we may need to find better solutions or alternatives.

You you enjoyed this article, please click the ❤ button below so others can read it here on Medium.