Troll Watch: On Eve Of Midterm Elections, Misinformation Targets Bigger Political Issues
MICHEL MARTIN, HOST:
This is where we've been trying to keep track of cybersecurity attacks, as well as the themes being pushed by bots and trolls. Well, Election Day has come and gone without any major allegations of coordinated attacks, but there were still plenty of false information shared on social media platforms leading up to the midterms. Kevin Roose has been monitoring that for The New York Times, where he is a technology columnist. And he's with us now. Kevin, welcome. Thanks for joining us.
KEVIN ROOSE: Thank you for having me.
MARTIN: You and your colleagues asked New York Times readers to send in examples of election-related misinformation that they saw online. You got something like 4,000 of them. What were some of the trends that you noticed?
ROOSE: Well, they ranged from sort of very low-tech misinformation mail that was sent to voters' houses up to text messages and Facebook ads and other examples of digital misinformation. One theme that we saw a lot was that instead of cropping up around specific candidates or specific races, misinformation would target these big national issues so things like the migrant caravan or the Kavanaugh confirmation hearings. Those would sort of create a big flood of hoaxes and doctored images and mislabeled videos and things like that that would spread on social media and would then, you know, be seen by hundreds of thousands, or even millions, of people.
MARTIN: So let's pick one, for example, the so-called migrant caravan. This is the group of mostly Central American migrants who've been traveling through Mexico toward the United States. You found several images shared on Facebook and Twitter claiming to be of this group. What were they actually?
ROOSE: Well, some of them were old. There was one post, for example, that got something like 36,000 shares, and it was a Facebook post that had three images of bloodied police officers. And this was being used as evidence that the caravan was becoming violent. And, actually, if you just do a reverse image search on those photos, you would find that they were old. They were, like, from 2012, 2014. They were from other protests, and they didn't have anything to do with the caravan. But they were put up as an example of what the caravan was doing.
ROOSE: I was going to ask you that. How did you trace the origins of these to demonstrate, really, almost immediately that many of these were false? How were you able to do that?
ROOSE: Well, one tool is called a reverse image search. You do it on a site called TinEye. And the way that it works is you sort of paste in the image, and it looks for previous instances of that image. And it'll let you sort of track the chronology of these images as they appear on the Internet.
MARTIN: You portray a very disturbing picture, frankly - you know? - of the landscape. You wrote that every time major political events dominated the news cycle, Facebook was overrun by hoaxers and conspiracy theorists who used the platform to sow discord, to spin falsehoods and stir up tribal anger. This is from a piece that you wrote. Two questions here. Are the social media companies - Facebook, specifically, but the others - doing enough to address this? And the second question is, what should the rest of us be doing.
ROOSE: Well, the companies are doing more than they have in the past. But we've seen that, as the companies start to defend against misinformation, the people spreading this misinformation actually adapt. They change their tactics. So one thing we've seen, recently, is that people will now take screenshots of text and share - instead of sharing text, they will take a photo of text and share that because it makes it harder for these automated systems to scan and flag misinformation if it's not there in text.
And I think the thing that users can do is just to beware. If there's no source attached to something, if you don't see a website that you recognize, if it's in a group or a page that you don't know, you know, what the origin of or how trustworthy it is, just hold off on sharing it maybe until you're a little more certain about the provenance and whether it's true or not.
MARTIN: Kevin Roose is a technology columnist for The New York Times. Kevin, thank you so much for talking to us.
ROOSE: Thank you for having me.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.