In the run-up to the election, fake news outperformed real news in engagement on Facebook:
Supposedly, Facebook is working on the problem, but there are good reasons to be skeptical—not just the financial incentives, but also the apparent naiveté there:
Google is also, quite smartly, working on cutting off the advertising dollars that fake news sites are generating with AdSense. However, Google is not the only ad network, so it will never be possible to completely stem the flow of ad money to these sites.
I definitely support the work the platforms are doing to address the fake news problem. But let’s not kid ourselves: they’re only going to do so much. Let’s not rely on the platforms to solve this problem for us.
I’m admittedly not an expert on the fake news problem, but I had an idea that I thought was worth putting out there: a bot that automatically replies to posts of known fake news. The bot would rely on a set of human-vetted URLs that are known to be fake news. It would then search for posts sharing those URLs, and go into action. It would reply simply, saying the post is fake news. When an independent fact-check is available, the bot could link to that.
How to compile this list of fake news URLs? Crowdsourcing. At first, a small team of people would submit URLs into the system to test it out. Once the initial bugs are worked out, it could be opened to a larger group. However, it will be important to do vetting, to make sure fake news publishers aren’t able to game the system. My modest proposal for naming this group is the Fact Brigade, but I’m sure there are all kinds of other awesome options.
This has some potential advantages over algorithmic approaches that some are discussing (Google is funding one):
- Each fake news story only has to be identified once. Once an item is identified, the system will engage any time it’s shared.
- The bot will gladly suffer the trolls, while you can do something more productive with your life.
- Algorithms are going to have false positives. Hopefully, a well-designed crowdsourced approach here could reduce that quite a bit.
One could even imagine a browser add-on that lets you report fake news even more easily. For example, Pocket and Buffer insert little buttons to save or share things you find on social media. How about a new button to report fake news?
There are already some browser add-ons that use lists of fake-news sites to display warnings, but I’m not aware of anything dedicated to reporting them.
This is, admittedly, only a very small piece of the puzzle. But I do think there’s value in calling bullshit, and doing it consistently.