Media Focus on Fake News: PBS Newshour Interview

LiT.RL News Verification Project receives mainstream media attention. pbs_fakenews_5_dec_2016Dr. Rubin was recently interviewed by the PBS Newshour Science (online print edition). Prior to that an op-ed article was commissioned by a former New York Times editor for the MIT Undark Magazine.






The PBS Interview ‘Uncut’

In a brief phone interview with the reporter from the PBS Newshour, Nsikan Akpan, we discussed the current fake news situation. Here is a fuller version of the information provided to Nsikan in response to his request for less prominent examples of the real world impact of fake news.

Nsikan Akpan: “For instance, has fake news influences other elections in the past? Or closed a school?”

Victoria Rubin: “Nowadays, the media at large refers generally to the problem of fake news but the distinction among the types of fakes have to be made. In one of our articles from 2015 we talk about 3 types of fakes:

#1. Journalistic fraud: here is an example in the U.S. context.

News, lies and videotape: The legitimation crisis in … Management’s response to these transgressions ranges from almost nothing — in Lang’s case — to a “cease and desist” for Mansbridge (and the rest of CBC’s …

Here is a Canadian example.

1012111-reportages-diffuses-plusieurs-medias-quebecois François Bugingo: des reportages inventés de toutes pièces … Le journaliste François Bugingo a inventé de toutes pièces plusieurs reportages internationaux. Lesquels sont vrais, lesquels sont faux? Notre enquête.

#2. Hoaxes. An example of a false tweet: “A powerful explosion heard from miles away happened at a chemical plant in Centerville, Louisiana #ColumbianChemicals,” a man named Jon Merritt tweeted.

The Agency – The New York Times The Agency. From a nondescript office building in St. Petersburg, Russia, an army of well-paid “trolls” has tried to wreak havoc all around the …

In the Canadian context, there was a minor example about a turmoiol in public schools in London, Ontario on Friday before the Canadian 2016 Thanksgiving. There were “clown attack threats” and police were sent to guard high schools. I’d qualify it as a hoax. Perhaps some mischeivious kids hoping to get an extra day added to the long week-end, but there was a huge turmoil in the city, since older kids were texting parents, and younger kids were panicking. The London Free press reported the incident here.

clown_lfp Clown threat sent to London high school students | The … Clown threat: Hoax reaches London, says clowns coming to London high schools to kidnap students, kill teachers

In my chapter in the Social Media Methods textbook, I describe rumor-busters or rumor detectors that deal with hoaxes. The abstract is here (in print to be published by February 2017). Some news outlets around the world such as Le Monde in France, employ these types of those technologies. #3. Satirical fake news can sometimes be mistaken for legitimate news by news reader or by journalists, as exemplified by the Trudeau’s “elbow gate”. Here is the actual incident video.

elbowgate_torontostar_18_may_2016 Here is the Beaverton’s satirical piece.

Entire NDP caucus arrive in neck braces, wheelchairs to … OTTAWA – All 44 members of the NDP caucus arrived in Parliament this morning wearing neck braces, arm slings, head bandages, and seated in wheelchairs after …

It was picked up, re-reported, and later retracted by the Hamilton Spectator. elbowgate_beaverton_19_may_2016

We have collected a database of legitimate news paired up with satirical fakes (based on content, by similar targets or topics of satire), and it was used for the creation of a successful implementation of our Satire Detector. Some news outlets around the world, such as Le Monde in France, have just started to employ some types of those technologies.

Here is an example of paired-up articles from the Health Sciences domain. A new drug, flibanserin or “female Viagra” has a large number of side-effects and low efficacy, was approved in 2015 by FDA. The lobbyists for Sprout Pharmaceuticals, the manufacturer, attempted to make the drug’s release an issue of gender-equality, despite the FDA’s concerns being focused on side-effects. The punchline from ‘a happy customer’, the Beaverton writes: “”With this new drug, my husband and I are intimate anywhere between one and one-and-a-half times more per month. Sure, I’m throwing up every three hours, but our love life is saved!” female_viagra_db

#4. The 4th type of misleading news is borderline deceptive: clickbait.

#5. And then there’s plain old peer-to-peer computer-mediated deception. We’ve written a lot about it since 2010. For instance, on discerning truth from deception with human judgments and automation efforts.Since then, we have developed a deception detection methodology based on its rhetorical features and successfully applied it to the news context .It was based on a dataset that we collected from the NPR’s Wait, Wait, Don’t Tell Me (with their permission). It identified lies in their Bluff the Listener segment from back in 2014. #6. The flibanserin example (and alike) led us into yet another type of fake analysis, the advertorials, editorials or news pieces that are sponsored by an advertiser, but whose content is provided by journalists, supposedly. This article from The New York Times is sponsored by Netflix’s Orange is the New Black is an example of barely visible ‘sponsored content’ labeling. One of my students, Sarah Corwell, is working on this particular variety of misleading news.

Women Inmates Separate But Not Equal (Paid Post by Netflix … Women Inmates: Why the Male Model Doesn’t Work. As the number of women inmates soars, so does the need for policies and programs that meet their needs

We don’t necessary follow how various fakes are propagated – we’re not into network analysis – but we use these example to motivate the need for critical thinking and the need for alert-type software to be developed, as you can see from the body of our work here at the LiT.RL lab at Western, Canada. We primarily focus on content-analysis of the news with an eye for automation with natural language processing and machine learning.”




Leave a Reply

Your email address will not be published. Required fields are marked *