All the news that isn’t
Researchers look at origins, spread of fake news on Twitter
Posted on: February 13, 2019; Updated on: February 13, 2019
By Craig Brandhorst, craigb1@mailbox.sc.edu, 803-777-3681
In the age of social media, it can be hard to tell truth from fiction. And when it comes to news — particularly if it’s related to a hot button political issue — fake news increasingly gets peddled as real news while real news gets maligned as fake.
Enter Mo Jang, an assistant professor in the School of Journalism and Mass Communications, and mass communication Ph.D. student Jo-Yun “Queenie” Li. Jang and Li were part of a research team that looked at the origins and spread of fake news on Twitter.
The researchers retrieved 307,738 tweets relating to 30 fake and 30 real news stories about the 2016 United States presidential election, then used a form of machine learning called evolution tree analysis to determine the root of each story and examine the story’s evolution online.
We wanted to use machine learning to help track fake news on social media. Hopefully, maybe social media companies can use something similar to help detect fake news before it spreads to the public.
Jo-Yun “Queenie” Li, Ph.D. candidate in mass communications
Their findings, which are published in the journal Computers in Human Behavior in July 2018, run counter to a common criticism of the news media made during and after the 2016 election.
“In the 2016 presidential election, news organizations were blamed for spreading fake news, but that is actually not true,” says Li. “We found that real news tends to be disseminated by news organizations but fake news is generated by ordinary people, by individuals.”
But that doesn’t mean fake news stories are always easy to recognize at first blush. In many cases, the root tweet linked to a non-credible website masquerading as a credible one.
“Often the original tweet links back to a website that may have a similar web address to a real news site but is actually a fake news website,” Li explains. “For example, it may have a name like ‘NBC News’ but the site is not actually related to NBC.”
Compared with real news stories, the fake news stories also spread much more quickly on Twitter but tend to flame out sooner than real news stories.
“Because fake news stories tend to originate from individuals, there’s usually not a legitimate website to link back to so the fake news is being spread one commenter at a time, and then very quickly their attention is pulled to other fake news stories,” says Li.
The team, which also included computer science and engineering professors Jijun Tang and Chin-Tser Huang and Ph.D. students Tieming Geng and Ruofan Xia, hopes to do a larger follow-up study and perhaps develop a machine learning tool that social media companies can employ in the fight against fake news.
“We wanted to use machine learning to help track fake news on social media,” she says. “Hopefully, maybe social media companies can use something similar to help detect fake news before it spreads to the public.”
Share this Story! Let friends in your social network know what you are reading about