5/27/22, 7:10 PM Faster facts: The rapid expansion of fact-checking – News in Asia Myanmar ICT for Development Organization (MIDO), which runs the IFCN-certified initiative Real or Not, teamed up with the country’s independent broadcaster Democratic Voice of Burma (DVB) to produce a media literacy show called ‘MILKyi’, while DVB also runs a separate fact-checking TV show with a variety of content. Many projects go beyond social media verification. Some focus on raising awareness about the potential risks and societal impact of misinformation, while others try unconventional ways to reach a wider audience. The ‘anti-hoax’ family drama series 28 by Indonesia’s Mafindo and Vietnam News Agency’s ‘No Fake News’ rap music video targeting youths, along with its ‘Factcheckvn’ TikTok project, 29 have gained large audiences through TV, video streaming services, YouTube and other platforms. ‘Factcheckvn’ won ‘Best Project for News Literacy’ in the Asian Digital Media Awards presented by the World Association of News Publishers (WAN-IFRA) in 2020. Automation via social listening and chatbots Newsrooms are also exploring the use of automation tools to flag content and engage with audiences in new ways. Fact-checking remains a highly manual task, where sifting through posts, images, videos, speeches and statements is a time-consuming first step to identifying potentially harmful misinformation. Newsrooms in Japan and Taiwan are putting their efforts into automating this process with dedicated fact-checking tools. In Taiwan, a tool using machine learning and artificial intelligence called ‘Rumor Catcher’ was built to help fact-checking initiatives that identify, collect, and cluster potentially harmful content in Chinese. Not-for-profit FactCheck Initiative Japan (FIJ) supports partners like national newspaper Mainichi and BuzzFeed Japan with three automatic fact-checking tools that its members can use. The Fact-Checking Console automates the monitoring and aggregation of potentially problematic content on the internet using machine learning and natural language processing. The ClaimMonitor then filters, tags and flags the content to notify media partners on Slack. When fact-checking stories are published by member organisations, the articles can be distributed through an app called FactCheck Navi that summarises the findings for the audience with links to the original content. 30 In Taiwan, a tool using machine learning and artificial intelligence called Rumor Catcher was built to help fact-checking initiatives that identify, collect and cluster potentially harmful content in Chinese. 31 The tool is still at an experimental stage, and its effectiveness in picking up misleading or false information worthy of investigation remains to be seen, according to a senior editor involved in the project. Building a system specific to a country’s language and social media environment, as seen with the efforts in Japan and Taiwan, is an attempt to supplement existing tools like CrowdTangle. It is a recurring talking point among fact-checkers in other countries in Asia, but most say they lack the resources and capacity to start such projects. Chatbots for messaging apps, on the other hand, have been implemented by many organisations across the region, including in India, Indonesia, Myanmar, Taiwan and Thailand on messaging services such as LINE, WhatsApp and Facebook Messenger. Typically, chatbots are linked to databases that store the details of past fact-checking investigations. Inquiries sent by users to the bot account are compared to claims in the database. When there is a match, the user will get an immediate response and links to the relevant stories. Different organisations take different approaches when there is no match, but often the user’s messages are sent to human factcheckers for further investigation. For this chatbot model to work, the database needs to have a sizable collection of datasets. For this reason, cross-organisational collaborations to share content and contribute to a single database are taking place in some countries. Most notably, in 2020 many fact-checking outlets in Asia participated in the #CoronaVirusFacts Alliance initiated by IFCN, contributing to a large database of more than 9,000 stories about Covid-19 misinformation. The database has become the backbone of IFCN’s chatbot on WhatsApp, which is available in multiple languages around the world. https://newsinasia.jninstitute.org/chapter/faster-facts-the-rapid-expansion-of-fact-checking/ 10/13