Key Findings
- From the Taiwanese presidential election on January 10 until the end of March, Doublethink Lab conducted a longitudinal observation of the posts made in 202 Facebook fan pages and public groups. During this time, 25% of posts (69,165 items) were related to Coronavirus. Based on the narrative trends of these items and the chronology of the development of the Coronavirus pandemic, it is possible to divide this period of analysis into four phases: “The virus is not severe,” “Point of outbreak and folk remedies,” “Taiwan’s government is covering up the epidemic; communities are already infected,” and “COVID-19 has caused disasters in several countries while the virus is already under control in China.”
- Of the 11 cases of narratives and themes that comprise this analysis, some of them bear clear traces of coordinated efforts. News items tended to “break” at the same time and using the same narrative structure as Chinese propaganda. Two examples can be found under the disinformation themes of “comparisons between Coronavirus and U.S. Influenza (H1N1)” and “various countries’ inability to control the epidemic.”
- The modes of dissemination used in the information operations built around “comparing the virus with U.S. Influenza ” resembled those of the 2018 “Guanxi Airport incident,” which originated with Weibo posts, whose content was then transmitted into content farm articles, then picked up by Chinese news media, and ultimately being quoted and published by Taiwanese media, whence it diffused to Taiwanese social media and forum websites. The difference lay in the fact that the earlier attack had its tone set by Chinese officials, falling under the “Outreach Mode” of the “four modes of CCP information operations;” this was not true for the COVID-related disinformation effort.
- Just like the 2019 “#vowing_my_intention_to_vote (#宣告我的投票意志)” disinformation campaign in Taiwan, the 2020 disinformation campaign around “#just_a_little (outbreak of)_pneumonia_in_the_Mainland (#大陸搞個小小肺炎)” qualifies as a “Collaboration Mode” attack, wherein purely economically motivated rent-seekers spread disinformation. The characteristics of “Collaboration Mode” are that disinformative content’s manufacture and distribution are carried out entirely by Taiwan-based actors and that the attack’s instigator and disseminators are separated, to the extent that it is impossible to identify the attacker behind the disinformation. Although the content used in these attacks is itself crude and easily identifiable, we anticipate that this mode will be further enhanced and become a primary method of Chinese cognitive domain attacks in the future.
- Compared to the earlier common pathways of disinformation dissemination, in which content was transmitted from Weibo to content farms, contemporary disinformation flows from videos posted on YouTube, which are then shared to Facebook or LINE. In addition to evading the rules of Facebook’s community guidelines, this transformation also allows attackers to take advantage of YouTube’s algorithms, which helps manufacturers of disinformative content optimize user engagement as well as to obstruct the efforts of detection and feedback systems. This represents a new form of “Content Farm Mode” attacks, which made up a large portion of disinformation in 2020.
- Although it is challenging to trace disinformation attacks back to a definitive attacker, the appearance of this content coincides with the same topics emerging as internal public discourse issues in China. From this, we can infer that much of China’s dissemination of disinformation is not purely for external propaganda purposes, but rather is intended as a domestic stabilizing social control measure. However, if a message or narrative is not widely spread within China, but is distributed widely within other countries, this would indicate that the information operation may not fall within the classification of a stabilization measure. Information operations of this sort should be given particular attention.
Key Findings
- Taiwan serves as one of the critical nodes in disseminating COVID-19 disinformation from China to Southeast Asia. China has taken advantage of the language similarity between two countries and sabotaged democracy in Taiwan; for example, our research has found that overseas Vietnamese in Taiwan and migrant workers are also targeted by disinformation — written in Vietnamese — to create chaos both in their hometown and Taiwan.
- Disinformation tinged with Chinese propaganda was shared by Chinese embassies’ Facebook fan pages in Romania and Slovakia and simultaneously shared by Myanmar and Vietnam Facebook accounts, later circulated into Southeast Asia. On the other hand, China state media — Xinhua, Global Times, and Reference News — would selectively debunk the disinformation that is against China's so-called “cheerleading” narrative.
- China-linked infodemic is not only “cheerleading” or “telling China’s story well” as previously assumed, but also negative and aggressive — harshly criticizing certain ideologies and fabricating conspiracies. Social media platforms have become hubs for effective dissemination of distorted content facilitated by WeChat, Weibo, and other social media.
- Overseas Chinese in Vietnam, Cambodia, Singapore, Malaysia, Myanmar, and the Philippines are influenced by WeChat official accounts while their Facebook groups and Twitter are used as hotbeds for disseminating disinformation.
- According to our research, the first spotted disseminators usually serve for China’s propaganda by echoing China state media, while the second disseminators dial down their pro CCP inclination by posting politically-irrelevant videos and posts. On the other hand, the consumers of disinformation — urged by the anxiety epidemic— also spread twisted messages into local communities.
