5 projects take different approaches to promote fact-checking, fight misinformation
This week a different kind of hack day took place at the MIT Media Lab.
These gatherings are usually filled with software developers and other technical folks. Wednesday's hack day had its share of geeks, but there were also social scientists, journalists, NGO workers and students from Harvard's Kennedy School, among others.
It was an interdisciplinary mix, and a nice complement to the Truthiness in Digital Media symposium held at Harvard the day before. (See this post from me about an interesting piece of data shared in one of the presentations.)
On Tuesday we discussed problems and outlined possible solutions. Wednesday's goal was to sketch out ideas that, even on a small scale, could attack some of the challenges.
The result was a variety of suggested projects that looked at misinformation in digital media. Each of the groups will post about its ideas on the event's blog, but Andrew Phelps at the Nieman Journalism Lab has already written a good overview.
We absolutely need more projects in this area, and I hope the half-day of activity at MIT can spark new efforts. At the same time, I came away with was an appreciation for the efforts already underway. Progress is happening.
Below is a quick breakdown of five projects focused on fact-checking and misinformation, all of which were represented in Cambridge this week. Even better, this week marked the first time they were all in the same place at the same time. By the end of the two days, there was a collective sense that they should work together as much as possible -- which I believe is the most important development of the event.
LazyTruth: I'll have more on this project when it fully launches soon, but the concept is simple and I think very useful: LazyTruth is a Gmail gadget that will alert you if you receive an email containing a fake chain letter, urban legend, or other type of documented misinformation. That way you won't be tempted to send money to that Nigerian prince. It also can nudge you to reply to the sender with the appropriate debunking information.
The trio of Matt Stempeck, Justin Nowell and Stefan Fox are working on LazyTruth. One aspect of this system that strikes me as particularly important is that email is frequently used by political misinformation campaigns, not to mention scams and other nefarious information-based attacks. Yet it's difficult to track the spread of these emails and to fight the misinformation -- one reason email is such an effective tactic. If LazyTruth can gain a good-sized user base, it can gather data about the spread of these campaigns. That's just as important as its ability to debunk misinformation.
TruthGoggles: Dan Schultz has already managed to attract a lot of interest for this project. Last year Nieman Lab did a great overview of what Schultz was trying to build as part of his masters thesis at the MIT Media Lab. (I often edited Schultz's posts when I was the managing editor of PBS MediaShift.) From there, the story spread to NPR and other outlets. I even heard him interviewed one Saturday afternoon on CBC Radio in Canada.
Why all the interest? Truth Goggles is a browser plugin that will tell you if something you're reading online includes claims or information that have been fact-checked. Let's say you're reading a story that refers to Mitt Romney's claim that Iran released its American hostages the same day Ronald Reagan was sworn in as president because of Reagan's "peace through strength" policy. Since PolitiFact rated that statement as "Pants on Fire," Truth Goggles would alert you that the statement had been checked and provide you with a link to the verdict. Schultz is due to graduate in May, so by then we should have something to look at and try out, though it will take more time to build out the full tool.
Fact Spreaders: This project is based at the University of Michigan and led by professor Paul Resnick. The concept is that more and more claims are being checked, but they aren't being spread to a large audience. So Resnick and his team are working to build a site and community that will help spread accurate information and verified claims.
The idea is to recruit a community of users who would reach out to people on Twitter and other social networks when they see them spreading misinformation. (For instance, a Fact Spreader could send a reply tweet explaining that the information someone just linked to has been debunked.) Resnick and his team also want to engage the crowd to help identify claims on social media and elsewhere that should be checked. That could act as a crowdsourced assignment desk for fact-checkers like PolitiFact.
PolitiFact API: All of the aforementioned projects require a database of fact checks. There's no way Truth Goggle or LazyTruth can operate if they don't have content to pull in and display to users. The good news, which I learned this week for the first time, is that PolitiFact has an API. This means their content can be accessed and used in applications (provided, of course, that thse applications adhere to the terms of service for the APIs).
On a related note, the hack day group that included Bill Adair of PolitiFact sketched out how fact-checks and verdicts could pop up onscreen when political campaign ads aired on interactive TV. That kind of application would also require access to the PolitiFact API; it shows how the such APIs could enable a new realm of applications for fact-checking.
Truthy: I previously wrote about this project at the University of Indiana, as did CNN. Its team is working to track how memes (and misinformation) spread on Twitter. They had success with an initial focus on political misinformation and astroturfing, and they're now expanding to other forms of memes and misinformation.
The Truthy website shows the diffusion networks of recent Twitter memes and describes how things spread on that network. They also invite people to flag suspicious memes. This crowdsourcing adds muscle to the system, which uses a "sophisticated combination of text and data mining, social network analysis, and complex networks models."
Need for collaboration
It's not hard to see how all of these projects could be working together. In fact, that was the focus of the hack day group I joined. We sketched out the basics of what a credibility API would look like, and how it could help feed the above projects and new ones. In a blog post about our efforts, LazyTruth's Stempeck explained why it made sense to enable everyone to share data:
There could be many benefits to working together in a federated network. Talented developers are prevented from experimenting freely in this space because of the high barrier to entry that is hiring a team of researchers. Fact-checking outlets can’t possibly know, empirically, which internet fire most needs a dousing. The audience for traditional fact-checkers is also limited to their relatively small web and print readership. Together, we could do much more, and at greater scale.
Amen to that. We need more projects and experiments, but let's also be sure to get these people to talk to each other and collaborate so that their tools can have the most impact.
Correction: This post originally stated that FactCheck.org has an API. In fact, they do not have an API for their content.