Fighting fake news isn’t just up to Facebook and Google
new york — You, too, can join the battle against misleading and other “fake” news online. But your options are somewhat limited unless you’re already an academic or data scientist who’s been studying the subject since way before Donald Trump started running for president.
Giovanni Luca Ciampaglia, a research scientist at Indiana University, fits that bill. He helped create a tool tracking how unsubstantiated claims spread online, a phenomenon that first caught his eye during the Ebola crisis in 2014.
“We started seeing a lot of content that was spreading, completely fabricated claims about importations of Ebola, (such as) entire towns in Texas being under quarantine,” he says. “What caught our attention was that these claims were created using names of publications that sounded like newspapers. And they were getting a lot of traction on social media.”
“Fake news,” which has gotten a lot of attention for its potential role in swaying the 2016 presidential election, has fascinated researchers for some time. Their studies have yielded tools that help track how “alternative facts” spread, and others that let you identify fake stories or block them altogether.
Deciphering Twitter rumours
Some of these are still baby steps in dealing with the phenomenon, but they’re part of a larger effort that now involves Facebook, Google and big media companies actively trying to tamp down the spread of fake stories. And the researchers were there first.
Tanushree Mitra, a doctoral student at the Georgia Institute of Technology, began a project three years ago to see how misinformation and fake news spread through Twitter. At the time, she says, “companies like Facebook and Twitter were not paying much attention.”
What attracted her to the project was the prevalence of fake news that spread online following natural disasters such as Superstorm Sandy in 2012. When she saw that people were sharing a lot of incorrect or misleading information about the events, Mitra decided to track both big stories and smaller rumours with the goal of creating an app that could help ordinary people sort fact from fiction so they can make decisions that could be crucial to their wellbeing.
Mitra and her fellow researchers scanned 66 million tweets linked to nearly 1,400 real-world events to identify words and phrases linked to perceived levels of credibility. Looking at tweets surrounding news events in 2014 and 2015 — including the Ebola crisis, the Charlie Hebdo attack in Paris and the death of Eric Garner in a confrontation of police officers in New York City — they asked people to judge tweets based on how credible they thought the posts were.
Words such as “eager,” “terrific” and “undeniable” were linked to more credible posts, while words such as “ha,” “grins” and “suspects” were the opposite. A computer matched the humans’ opinions 68 per cent of the time. The next step, an app, could help people rate the credibility of tweets and other social media posts.
A group of researchers at Indiana University have created an online tool called Hoaxy that seeks to visualise “the spread of claims and related fact checking online.”
Although it’s still a work in progress, Hoaxy can trace the origin of, for instance, the false claim that millions of votes in the 2016 presidential election were cast by “illegal aliens.” Type in your search terms and Hoaxy will report back with stories that spread the claims, as well as fact-checking articles that debunked it.
In this instance, the claim goes back to a November article from Infowars.com that was shared 17,961 times on Twitter and 52,200 times on Facebook, according to Hoaxy. The site only tracks actual links people shared, so it misses anything that’s paraphrased
What caught our attention was that these claims were created using names of publications that sounded like newspapers. And they were getting a lot of traction on social media Giovanni Luca Ciampaglia, Research scientist at Indiana University
or posted without a link.
A data visualisation tool shows the intertwined web of Twitter users who spread both the claims and the fact checks, and how they are connected to one another. The researchers focused on Twitter because the service makes more data available to the public, which makes it easier to use in data-tracking tools than Facebook. Tools like Hoaxy or rumour-identification apps are only helpful if people use them. The same goes for another approach — using a web browser plug-in to identify or block fake-news stories. For instance, the Chrome extension “Fake News Alert,” created last year, says it will tell you when you are visiting a site “known for spreading fake news.”
But there are a few drawbacks. Many people aren’t willing to go to the trouble of adding new extensions to their browser. And such extensions only work on the desktop version of Chrome, not its mobile counterpart.
“Fake News Alert” also uses a widely circulated but oft-criticised list of fake and misleading news sites assembled by a Merrimack College professor. The list casts a very broad net and includes some established, but highly partisan sites such as the right-wing Breitbart News and the left-wing Occupy Democrats. —