In case you missed it, Fake News is the topic of the month. On the tail of the biggest political upset of the century, Facebook and Twitter have come under fire for allowing fake news websites to spread without resistance across social media, often gaining much more attention than real news stories. Always gaining a lot more attention than the corrections which follow them.
One of the fake news authors recently admitted to Caitlin Dewey of the Washington Post that he does believe that he helped get Donald Trump elected with some of his run away viral lies. Craig Silverman and Lawrence Alexander of Buzzfeed news recently discovered the 100+ pro-Trump websites in the Balkans which intentionally publish a stream of the most inflammatory headlines they can create to attract traffic to their advert-covered-websites. The more outrageous their headlines, the more viral it went, the more traffic they would receive and the more money they would make.
Apparently Facebook didn’t do anything to stem the spread of this misinformation because they had already received criticism for suppressing pro-right articles from their trending topics feed. Since then Mark Zuckerberg has come out and said (while eating a dolphin, according to this article I found on Facebook) that they are actively working on a better solution now, though he still doesn’t believe that fake news had any real impact on the election.
Rise of the Fake News Plugins
In response to all of this media attention, and the genuine outrage over this serious problem, many people have started to work on external solutions to the problem. Business Insider reported that “It only took 36 hours for these students to solve Facebook’s fake news problem“. Those students invented FiB – Let’s stop living a lie, a browser plugin which uses AI to verify claims, and then correct them where possible if they are false. Lifehacker reported on the B.S. Detector plugin, saying “B.S. Detector Lets You Know When You’re Reading a Fake News Source“. This plugin, developed in under 1 hour, simply checks Facebook posts against a list of known Fake news websites and flags posts which are from them. Using the same technique, Brian Feldmen developed Fake News Alert. Esquire reported on that “This Chrome Extension Will Alert You If a News Site Is Fake“.
This is all so great to see. People feeling passionate about this issue and investing their free time into solving it is an amazingly positive thing. I hope we continue to hear about more of them in the weeks to come. This is a huge problem and needs a lot of people working on it.
But this is where I have to offer my experience to temper the excitement.
The web doesn’t actually need more plugins in this space. It doesn’t need more new projects trying to solve a problem which has actually existed in one form or another since the Internet became popular.
What we really need is a unified effort working on one of the many well established projects. We need to put that mental muscle behind something which has already carved a path forward and demonstrated its viability.
We need at least one wildly successful solution and not another 10 or 20 false starts adding to the already cluttered space of web annotation and fact checking plugins which have come and gone over the last 20+ years of the web.
Why Fact Checking Plugins Fail
As much as I appreciate the efforts of these new plugin developers, there is a harsh reality about developing in this space.
Fact Checking plugins aren’t cool.
People don’t enjoy being told they are wrong. And as we can see with the growth of the fake news phenomenon, most people don’t even particularly care if something is true or not. Silly pointless fun, good feelings, and sharing with friends and family are popular. Error checking news article is a bit too much like schoolwork, and most aren’t interested.
So the idea that making a plugin in this space will result in some run away success and rapid growth as everyone gratefully goes out of their way to install a fact checking / fake news alerting plugin, then proceeds to sing its praises and tell everyone they know, is unfortunately very unlikely.
But that is just one small part of the problem. The bigger problem is that there is no money here.
Making fake headlines to get views on your website allows you to make money through advertising. Warning people to not click on links is the exact opposite of how you make money. There is almost no way to fund this sort of project, so you need a pretty serious commitment to guarantee that your plugin will continue to be accurate and functional for more than 3 months of excitement after the free publicity and press.
Because, unfortunately, interest will be lost.
Right now, everyone is talking about this. Articles are being published and shared all over the place. Each of these apps have shown rapid growth with thousands of new users in only a week. But that will stop happening over the coming weeks (maybe even days?) and as soon as that stops, how do you continue to drive growth of an app which has no content creation and no viral component? You don’t. And so, like the many who have come before, these new efforts are likely to hit the same walls. No practical growth method and no money to justify the ongoing time and energy requirements, leading to the inevitable decline and death of the project.
Survival – My experience with rbutr
I know about this, because I have been through it all. I have hit those walls. I have experienced the difficulty of growing an app which has no viral component, creates no content, and serves a vital but essentially ‘uncool’ function. I have struggled for years trying to figure out how best to proceed with no money.
rbutr has managed to survive without money for almost 5 years now, but not without long periods of inactivity and a growing list of bugs and frustrations.
Those five years have given me a lot of time to think about how to solve these problems. I still don’t know the ideal solution to the funding problem, but I have chosen a path forward and outlined it recently. As for the problem of misinformation, scams and fake news, well, this is the area I feel most confident on.
The need for Integration
The key is to build a platform which can easily be accessed by other companies (like Facebook, Google, Twitter, etc) and directly integrated into their system under their own terms.
As long as people need to install a plugin to warn them about fake news, then we’re not likely to ever reach the people who really need to be warned about it. The reality is that only the most tech savvy people install browser extensions, and tech savvy people are more likely to be aware of fake news websites already.
So no, the only way to stop fake news, is to stop it at its source.
What this fake news phenomenon has clearly revealed is that more and more people are just reading headlines on Facebook and Twitter, and then sharing them from that snippet of information. People (for the most part) aren’t installing plugins, and now they aren’t even visiting the pages we’re trying to fight.
So the only way forward is for Twitter, Facebook, Google and other such companies which act as information portals to take the pro-active role here and give built in “quality assurance” of information. They have to do it, or else it simply won’t work.
But that presents a massive complication for them: Who wants to trust corporations with determining Truth for the public?
Confirmation Bias – The real problem with filtering misinformation
As soon as you publicly judge the accuracy of information, you open yourself up to backlash from people who disagree with your assessment.
This is the big problem with all efforts to fact check and offer ‘unbiased’ information. No matter how well you do the job, there will be a percentage of the population which holds beliefs contrary to your ‘factual conclusion’. Those people are far more likely to ignore your conclusions, judging them as biased and not take the information on board and change their beliefs. This is a huge problem for Facebook and Google who must be seen to be impartial.
Facebook had human moderation of their trending topics until a conservative backlash made them stop. They were accused of suppressing pro-conservative articles. What if all of the articles which they supposedly suppressed were actually fake news articles which just happened to be fake in the conservative favour? Apparently, that doesn’t matter; someone claimed they were bias in the media, and Facebook shut it down.
The second you step into the world of judging the accuracy of information, you open yourself up to backlashes. You open yourself up to loss of market-share as people boycott your service out of disgust with your “partisan” and “manipulative” filtering.
The solution: Outsource the critiquing and make it universal.
Basically, you have to let someone else take the fall for being “biased” while you, in full neutrality, simply make it possible to access third party critiques, corrections, and rebuttals. If the information is made up or a scam, the rebuttal will quickly reveal that. Where the information is controversial, the critique will simply provide more illumination to the debate, and people may come away from it better informed. Where the information is true, the critique will likely persuade no one, leaving everyone already believing what they already believed anyway.
In all cases, people will make up their own minds anyway. This is simply a truism. But when every piece of content provides easy access to the best criticism of itself, the world is more reliably informed and better primed for critical thinking. And that is a huge improvement over the current situation.
This is obviously how rbutr works, and I believe in this approach completely, otherwise I wouldn’t be dedicating all of my time to it. But I also think it is the only solution which can practically work at this level for this problem.
That said, the Fake News website phenomenon has made me aware of another feature which we can easily add to rbutr to have a wider impact.
The domain level rebuttal
Historically, rbutr has only had webpage level rebuttals. It makes the most sense in general since most arguments are made on single pages. But the domain level rebuttal is a simple way to allow articles to critique the overall reliability of websites. This is useful for satirical websites, fake news websites, and for a more detailed critique of non-fake news sources (none of which are perfect).
The idea here, is that critiques of a website, say Fox News, could be added to their homepage via rbutr, and voted up in the usual way. But the rebuttals to the homepage of the website wouldn’t fire like normal rebuttals – they will instead simply be accessible from the rbutr plugin on any page within the Fox News domain.
For Facebook’s purposes, the same value should be provided in their news feed. Every post which links to any article from a critiqued domain should provide an indicator of some sort which allows easy access to those critiques. It should be subtle and non-invasive, but clearly present and accessible from any view of the shared link.
In the end, almost every single website would end up having one of these alerts. Everything has detractors. But this state of affairs is far superior to the current situation where everything is presented as if it is true and with the backing of social echo chambers.
I also want to add that I think the domain level criticisms should be graded to some extent, because “Satirical” websites are intentionally lying for the sake of humour, and can be uncontroversially tagged as such. These websites should probably be more clearly indicated as “uncontroversially not true” in some way. See my last blog post about the Fake News label for more information on this topic.
I don’t think you could really take it further than that though without falling straight back into the problem of confirmation bias and backlash territory.
Calling out a website as fake news or untrustworthy is rife with complications and negative consequences. The best solution is to give easy access to third party critiques of each page and/or the whole domain, and let the users decide for themselves what they think. This approach needs to be implemented within browsers universally, but also needs to be implemented onpage by companies like Facebook, Twitter, and even Google within their search results.
This is the key solution to the problem of fake news, misinformation and false beliefs. Every other approach works to further divide people into their personal confirmation bias bubbles and increases the problem.