Ahead of Brazil’s Oct. 7 presidential election, the country’s fact-checkers are working around the clock to investigate questionable claims and debunk misinformation. But as soon as journalists uncover the truth about a bogus story, they face the next challenge: making sure that the readers spreading the story actually find out it’s a fake.
Enter Fátima.
The Twitter bot, whose name is short for Fact Machine, searches for tweets that link to discredited stories online. When it finds one, it automatically replies with a short message and a link to verified information. Most people who have been alerted by Fátima this way click the link, and some have also corrected their posts. A few have even thanked the bot.
Muito obrigado, vou verificar!
— Ricardo di Paula (@Ricdipaula) July 28, 2018
Election season is the most critical time to correct fake news, according to Tai Nalon, one of the brains behind the bot. Nalon is founder and editor-in-chief of Aos Fatos, a Brazilian media startup dedicated to fact-checking. “We have to reach more people in order for fact-checking to be more efficient and effective,” she said.
Fátima is powered by a database — continually updated by Aos Fatos fact checkers — of links to misinformation, and corresponding links that debunk the false statements. The bot, which scans Twitter hourly, helps Aos Fatos’ journalists transcend filter bubbles and reach people exactly when it's most needed: when they are sharing misinformation that has already been debunked.
More than 9,500 Twitter users, including public figures in media and politics, follow Fátima. Deltan Dallagnol, lead prosecutor of Operation Car Wash, a now four-year-long corruption investigation into the Brazilian oil company Petrobras, recommended the bot to his followers.
To date, the bot has corrected the senders of more than 1,800 tweets, prioritizing Twitter users with high numbers of followers and original tweets.
Nalon and Pedro Burgos, an International Center for Journalists (ICFJ) Knight Fellow whose tech team is supported by the Google News Initiative, worked together to build the bot at Aos Fatos. The week before the election, they also launched a Facebook Messenger version of Fátima that gives people tips on how to evaluate the accuracy of the news, images and videos they encounter on the internet.
@goesjorge1 Olá! Não é verdade que Boulos liderou as buscas feitas no Google durante o debate da Band. Ajude a espalhar a informação correta: https://t.co/8toZvc09qO
— Fátima (@fatimabot) August 15, 2018
Burgos first had the idea of building a fact-checking bot in 2016 while working as an engagement editor at The Marshall Project, a nonprofit news outlet. During coverage of the U.S. presidential debates, Burgos anticipated that then-candidate Donald Trump might provide incorrect information about crime statistics so he prepared fact-checking tweets in advance.
“The fact that I tweeted those charts just seconds after Trump talked about the issue made a huge difference in terms of engagement, as people were more interested in the topic,” Burgos said. “That’s the logic behind all successful recommendation algorithms from the Facebook newsfeed to Netflix: they suggest something relevant to you in the moment you’re interested in looking for something.”
Oi! Esse link sobre LGBTs e pedófilos que você compartilhou é de uma notícia falsa. Ajude a espalhar a informação correta aqui: https://t.co/MSQ2hYP3s7
— Fátima (@fatimabot) August 1, 2018
Aos Fatos plans to add more features to Fátima so the automated bot can check facts in real time, or close to it. The project recently won a flash grant to expand from the International Fact-Checking Network. The bot currently replies to links sharing misinformation but it could potentially also evaluate the validity of text, images and videos on Twitter and other platforms.
It’s an ambitious project, but Nalon said a tech-oriented philosophy is central to Aos Fatos’ journalism.
“My advice for other newsrooms wanting to automate their processes is go for it, but collaborate,” Nalon said. “Innovation is expensive, and there’s not enough money for everyone out there, especially in Brazil where we are recovering from a very deep financial crisis.”
While the spread of false information is a chronic problem, she said, “We can do more, and we can do it together.”
To see the bot in action, follow Fátima on Twitter: @fatimabot.
Main image courtesy of Fátima's twitter page.