In Taiwan’s four-year presidential election, candidates work hard to campaign, there were a lot of news produced every day, but no one could be sure whether these information were true or false. Starting in September 2019, we turned the public conversations of presidential candidates into texts in crowd-sourcing way, cooperated with 12 media and research institutions to check. We also provide real-time fact checking at the only presidential debate.
In the 4 months since the project, at least 1,432 netizens have worked together to provide the media with conversation texts for fact-checking In the end, the media completed 230 fact-checks, found 47 wrong messages, 60 one-sided facts, and produced 15 in-depth reports.
We also provided real-time fact-checking during the presidential election debate. The 6 media partner immediately check the candidates’ conversations and complete 46 fact-checks within 8 hours. This is the first cross-media and cross-field cooperation project in Taiwan. It was done by the media and netizens in cooperation.
We collect candidate speech videos and divide them into one video in units of 15 seconds. Upload them to youtube to generate video links, and provide google forms to volunteers to convert videos into text. And we use google forms to design a mechanism so that volunteers can verify the correctness of the text. The result will be stored in google sheet. We use Vue.js and Google Sheet API to implement web page.
What was the hardest part of this project?
We split the daily public conversations of presidential candidates provided by the TV station into 20-second-unit videos, and handed them to netizens to transit into text. Trained volunteers marked these messaeges that need to be checked in the conversation,and then these messaeges could finally reach the media for fact-checking. We built a collaborative platform from scratch. Since we hope to start the project as soon as possible to give voters as much correct information as possible before the polling day, we used Google spreadsheet to build all processes.
Cross-media cooperation was also a difficult thing. We must find like-minded people, established the procedures and standards for fact checking, not to mention the pursuit of exclusive news by many media. All these make this large-scale cooperation project valuable.
What can others learn from this project?
Perhaps a small team can do limited things, but combining the power of everyone can make a big project.