“An earlier version of this column was published in error. That version included what purported to be an interview that Kanye West gave to a Chicago radio station in which he compared his own derrière to that of his wife, Kim Kardashian. Mr. West’s quotes were taken, without attribution, from the satirical website The Daily Currant. There is no radio station WGYN in Chicago; the interview was fictitious, and should not have been included in the column.”
This is an actual correction that was issued by our nation’s most trusted rag, The New York Times, in November 2014. Let’s face it: there is simply no comparison between the Kardashian and West derrières, even in the ego-addled mind of the latter. That an error this glaring could sneak by a team of world-class fact-checkers, researchers, and editors just goes to show how rampant and insidious bad information has become in our media. Mistakes are often amusing, sometimes damaging, but regardless of their gravity, they have the potential to cause you to be wrong. Whether you’re a writer or reader, a publisher, a brand or even a child on a playground with a reputation for credibility, this is something you need to care about.
The explosion of digital journalism and online publishing is obviously to blame here. While corrections are easier to make, it’s also easier to make errors, which spread by social media, are as infectious as the flu in a classroom. It’s hard to quantify the effects, but we are producing and consuming so much more information than we used to on a daily basis that we have exponentially less time to check—perhaps even to care if it is right or not. As Bill Adair, the founder of Politifact told the American Journalism Review (API) earlier this year:
“The information age is … wonderful for the breadth of information that people receive, but it’s also a time when people need to sort out what’s true and what’s not…it’s never been easier for politicians to spread falsehoods than it is today and so it’s critical that journalists fact check what the politicians are saying.”
In 2014, the American Press Institute’s fact-checking project began researching research methods and found that “political misinformation on Twitter outnumbers the people trying to fight the misinformation 2.7 to 1.”
That means 3 in every 4 falsehoods tweeted out go uncorrected. Which is crazy. Can we really afford to waste so much time learning lies?
[Tweet “3 in every 4 falsehoods tweeted out go uncorrected. Can we afford to waste time learning lies?”]
Can technology help us stop the infectious spread of digital misinformation?
For those of us who care about actually being right, and doing more than just spewing out half-truths garnished with opinion, it should be of some comfort that fact-checking is growing around the world. In fact, the 2014 fact-checking census from the Duke Reporters’ Lab found a 50% global increase in fact-checking sites.
And while reputable news outlets and specialty projects have been putting their manpower behind sites like FactCheck.org, PolitiFact, and the Baloney Meter, apolitical data scientists are now pitching in, developing automated methods to verify facts.
Scientists at the University of Indiana have developed an algorithm that can leverage a body of knowledge (like Wikipedia) to help fact check a document. A paper this June in the journal PLoS One, explained how the program assigns “truth scores” to statements about history, geography, and entertainment, and is able to answer simple questions. For these preliminary experiments, the scientists built a “knowledge graph” from the infoboxes on Wikipedia. Using 3 million concepts and 23 million links between those concepts, the algorithm can decide what is “most factual,” based on that network of interconnected facts.
[Tweet “The technology “could function like an advanced spell-checker,” but for facts” “]
“While the algorithm is several years away from being applied to an actual technology, it “could function like an extremely advanced spell-checker, in Word” but for facts, said Giovanni Luca Ciampaglia, a postdoctoral fellow who led the study in an interview with Oz last week.
Ciampaglia, whose research is about how information is created and consumed, was motivated by the same problem that’s been driving the growth of fact-checking operations worldwide. Although there is an enormous volume of information being generated online, there are not enough resources focused on defining true, quality content and good information. Ciampaglia echoes Adair and API in the urgency of this mission, which causes us to wonder: what sort of critical missions get taken on without a little help from technology?
According to Ciampaglia, there is very little help for journalists performing the fact checking and it’s almost all done manually. But don’t abandon your masters degree in research just yet, human fact-checkers won’t be replaced, just helped, “my impression is it’s not going to replace people,” he laughed, but rather, he explained, computational fact-checking can help journalists keep up with the enormous volume of information generated online and enhance our ability to evaluate the veracity of dubious claims.
An algo that can help you call others on their BS and make sure you are correct? That sounds right.