Nieman Foundation at Harvard
HOME
          
LATEST STORY
From shrimp Jesus to fake self-portraits, AI-generated images have become the latest form of social media spam
ABOUT                    SUBSCRIBE
Nov. 30, 2018, 12:16 p.m.
Reporting & Production
LINK: medium.com  ➚   |   Posted by: Christine Schmidt   |   November 30, 2018

WhatsApp blasts past other platforms in terms of immediate information sharing. But until recently, no news organization had access into the “black box” of misinformation’s backend.

My colleague Laura Hazard Owen has chronicled WhatsApp’s impact in misinformation and news distribution abroad, especially as elections in Brazil, Mexico, and elsewhere translate to closer scrutiny of information flows. Remember, WhatsApp is owned by Facebook, which is still trying to figure out how to monetize its more than 1.5 billion monthly users worldwide.

While Facebook works on that, news organizations have struggled to counter its spread — the BBC, for one, recently conducted in-depth research on WhatsApp’s virality in Kenya, and Nigeria. News outlets also team up around elections in various countries in attempts to tackle election-related misinformation like, you know, what happened in the U.S. Verificado 2018 was a a collaboration between Animal Politico, AJ+ Espanol, and Pop-Up Newsroom for Mexico’s July general election. But it’s still a big lift:

Verificado’s WhatsApp line officially launched on May 18; two weeks in, 4,800 people have subscribed to it and it has received a whopping 18,500 messages — 13,800 of which it has answered. (Maccise cautioned that “messages” here refers simply to interactions — requests to verify content, as well as any other message, such as a greeting. But Verificado plans to filter out actual fact-check requests more specifically in a report it will release after the election.) Each of its published debunk statuses has gotten about 1,000 views (though in reality it’s probably more, since users who disable the read receipt feature aren’t counted in WhatsApp’s analytics). And all of this is handled by just 4 people on Verificado’s end.

In Brazil, a collaboration among 24 newsrooms had a secret weapon: Better infrastructure thanks to special WhatsApp access. Comprova, a verification project from Claire Wardle et al’s First Draft effort against misinformation, gained access to WhatsApp’s API (not to the content privately shared on the platform) to run its fact-checking operation around Brazil’s presidential election. Here’s how they set that up, by Aimee Rinehart:

For the first time, WhatsApp gave a journalism non-profit the ability to receive questions from the public on an intermediary website, Zendesk. Had we just used WhatsApp for Business, we would have been tethered to one phone to try and respond to the almost 70,000 questions the project received. Zendesk was integrated into our business account and allowed multiple journalists and student journalists to review “tickets,” or questions from the public. For most of the project, four journalists monitored questions. For the last month of the election, there were 10 people assigned to sort and sift tickets. Through this process, the team began to see patterns of mis- and disinformation and sent these on to editors for consideration to debunk. The team also replied to the person who sent the original message with a simple note or a link to the published article on the topic.

First Draft prepared for Comprova based on its experience from organizing CrossCheck France last year:

The collaborative project CrossCheck France ran for 10 weeks in spring 2017 and received 600 questions from the public by email from a submission form on the website. We were warned by Brazilian journalists that we could expect to see 600 questions in the first hour from Brazilians. They said so without a laugh and they were right about an engaged citizenry: Comprova received almost 70,000 questions from the public. During the first week of October leading up to the first election, the business account received an average of 3,000 questions daily. The Zendesk project manager reported that the platform typically receives 2,000 questions per month for its most-active clients, so Comprova struck them as a particularly “viral” experience.

Now, First Draft with the International Center for Investigative Reporting has launched CrossCheck Nigeria with 16 news outlet partners ahead of the country’s 2019 general election, where WhatsApp is the most popular messaging app.

Show tags
 
Join the 60,000 who get the freshest future-of-journalism news in our daily email.
From shrimp Jesus to fake self-portraits, AI-generated images have become the latest form of social media spam
Within days of visiting the pages — and without commenting on, liking, or following any of the material — Facebook’s algorithm recommended reams of other AI-generated content.
What journalists and independent creators can learn from each other
“The question is not about the topics but how you approach the topics.”
Deepfake detection improves when using algorithms that are more aware of demographic diversity
“Our research addresses deepfake detection algorithms’ fairness, rather than just attempting to balance the data. It offers a new approach to algorithm design that considers demographic fairness as a core aspect.”