Real-time Fact-checking
We all, more or less, know what the term “fact-checking” means but may not be familiar with what “real-time fact-checking” is all about.
“Autocorrect: How advances in real-time fact-checking might improve our Politics” by Jonathan Rauch (Atlantic magazine June 2019) describes a system, Autocorrect, designed by Bill Adair and his team at Duke University.
Here is a brief example of how the Autocorrect system would work:
Assume that your TV or Cellphone has the has the Autocorrect app installed and is executing in the background while you’re watching a political speech. The app will intercept the broadcast signal and delay it for 30 seconds while it searches the web analyzing the content of what the speaker has just claimed. If the claim has already been verified or debunked by an independent fact-checking organization, the app returns one of three possible verdicts: “True, False, or Not the Whole Story” plus any other useful information found. As the author concludes, If the system works you will see the verdict as the politico speaks --- “no waiting for post-speech reportage, no mental note to Google it later. All in seconds,without human intervention. If it works.” Pretty impressive huh?
Contrast this with the current state of fact-checking which is to enter the claim into your favorite search engine and sit back waiting for the listing of the hits on your query. This method has many faults, the most obvious one being --- who’s got the time to visit all of the sites returned?
However, this raises the larger and more important question,” Whom do you trust? “ How do you decide which of the numerous websites are legitimate or (in a worst case scenario) just trying to sell you a time-share in the Bahamas or change your political stance?
In an attempt to deal with the legitimacy of search results, I used my search engine to find, “sites that review the reviewer sites”.There were a few and the most interesting one (to me) was a review of Amazon product reviews called FakeSpot. According to the app, “ Fakespot provides consumers with a new way of filtering product reviews to find out what real users are saying about the products you want to buy. Our proprietary technology analyzes millions of product reviews, looking for suspicious patterns and incentivized reviews. We then weed out the reviews we think are unreliable and return a grade of A,B,C,D or F which grades the the authenticity of any particular review. But you may ask, “who reviews the veracity of these supposedly unbiased reviews of the reviews? Where does this investigation end? Whom do I trust?
I think trust is usually based on a chain of beliefs. For example, I trust The Atlantic Magazine which trusts the author of this article and Rauch trusts Duke University which in turn trusts Bill Adair, the head of the group which creatPolitifact. So by following this chain of trust, I trust Adair’s claims. Now if you ask me on what basis I trust the Atlantic Magazine, that’s a tough one. The best I can do is say that it’s been around for over 150 years and its articles compared to newspapers and tv news are longer allowing them to pursue the issues to a greater depth. Also important is that the Atlantic has stood the test of time and I think that the longer a news medium has existed, the lower the odds of its reporting fake news thus the higher its credibility.
So my final question is simply, “Is this a good or a bad thing? It certainly won’t increase our delayed gratification quotient that a successful adult should cultivate. But it will increase our usage and dependency on the Internet further entwining it’s good and bad qualities into our everyday life. Only time will tell or as our current President has replied on many occasions to tough questions from reporters, “We’ll see.”
You can find podcasts describing Adair’s work on Autocorrect at:
