The Chinese Infodemic in Taiwan
A Preliminary Study on the Dissemination Model of Disinformation, Taking COVID-19 as an Example
Author: Poyu Tseng, Puma Shen
Translator: Shih-Shiuan Wilson Kao, Stanley Christopher
Editor: Lilly Min-Chen Lee
This research was conducted in May 2020 and the data was collected between Jan 10 to Mar 31, originally written in Chinese.
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.
1. Research Purpose
Messages containing disinformation are disseminated more quickly in the digital era through the advancement of information technology and infrastructure and have become far more destructive to democracy. “Fake news” and “disinformation” have been trending keywords in the news report. However, most of the people, influenced by confirmation bias, still tend to believe statements that agree with their ideologies or what they want to believe, not statements based on facts and shreds of evidence. Later corrections and clarification don’t help to remove such misperception but strengthening their current beliefs¹.
Large-scale information operations by disinformation had been observed before the 2018 local election in Taiwan, and “information warfare targeting Taiwan that initiated by China” has been brought to attention. However, though most people do recognize “fake news” could influence the election, researches on information warfare targeting Taiwan are far from comprehensive, and collection of relating evidence remains difficult; therefore dissemination tactics and impacts of disinformation in Taiwan are not well-understood by the civil society. Such gap of understanding leads to cognitive biases such as “opinions that are ideologically different from mine are all disinformation”, therefore “information operations in Taiwan” may become merely a mouthpiece in the middle ground of politics rather than a serious issue of social trust and national security.
To further understand the dissemination pathways and tactics of disinformation, we had monitored and archived messages online, grouped them by time, themes, and narrative frames, and performed in-depth text analysis.
2. Research Methods and Framework
We had monitored the most significant in recent trending topics, that is COVID-19 (commonly referred to as “Wuhan Pneumonia/武漢肺炎” in Taiwan,) between Jan. 11 — Mar. 31, 2020. This time range can be further divided into four phases with respects to different trending narrative modes:
We had monitored and archived posts from 202 suspicious Facebook fan pages and public Facebook groups. Along with this archival, data described in the report published by Taiwan FactCheck Center and MyGoPen, and data inside of rumor report/verification for the LINE database by Cofact were also included in the dataset of this analysis. 69,165 posts were filtered from 269,305 posts, ranging from Jan. 13 to Mar. 31. After the removal of purely journalistic messages, we further drew 8 themes and 13 narrative frames (see Table 2 to Table 4.) Themes are the topics of discussions, i.e., what people talk about; presuppositions of discussions are given by narrative frames, by which people can express their opinions and perspectives. Each message may contain multiple themes and narrative frames.
The analytical process involved first preparing case studies for each theme and sorting narrative frameworks chronologically, including highly detailed contextual explanations of these frameworks to aid observation of trends in the dissemination of information. Finally, we compared the frameworks and dissemination trends for similarities and differences.
3. Results
(1) Case Studies
Phase 1: Jan. 11 — Jan. 23, 2020 (3 themes and 5 narratives)
COVID-19 began to spread in China long before the presidential election voting in Taiwan (Jan. 11). News reports were scarce in the beginning since no large-scale outbreak was reported, and public attention was being drawn to the results of the election and upcoming political affairs. The first confirmed case in Taiwan was reported in Jan. 21 and news reports began to surge; at the same day, numerous related posts appeared on social platforms such as Facebook, PTT, and DCard; president, premier and all 6 mayors of municipalities of Taiwan all posted on their Facebook pages telling the public to pay attention to the outbreak but need not overreact. Wuhan authorities declared travel restriction on Jan. 23, referred to as “Wuhan lockdown” in Taiwanese media, but the number of news reports dropped slightly since it was Chinese New Year’s Eve that day.
Case 1: Beware contracting pneumonia while voting
Summary: Many messages appear just before the day of the presidential voting day (Jan. 11), reminding people to wear masks during voting to prevent catching the unknown Wuhan pneumonia, and there had already been confirmed cases in Taiwan², suggesting that “voting is risky.” These “friendly reminders” contained disinformation of “confirmed cases in Taiwan” and may incite people’s panic to suppress voting.
Background information: There had been intelligence that “pneumonia outbreak in Wuhan is similar to SARS.” Taiwan began to inspect flight departed from Wuhan on board since Dec. 31, 2019, and Taiwan CDC enlisted “Severe Specific Communicable Pneumonia (嚴重特殊傳染性肺炎)”³ as a category 5 communicable disease since Jan. 15⁴. No confirmed cases had been reported at that time.
Coverage: These messages were widely circulated on LINE and Facebook. There were ten user-filed reports in Cofacts (a rumor report/verification initiatives) database (Two analyses of user-filed reports on Cofacts: the first one, the second one)⁵ for LINE. The Taiwan FactCheck Center also issued a report one day later.
Follow-up: The fact check report was issued within a single day, with timely clarification by Taiwan CDC that “there have been no confirmed cases in Taiwan yet.” Since this message expired soon and had been debunked quickly, these messages were not observed circulating after the voting day.
Case 2: “Just a little (outbreak of) pneumonia in the mainland.”
Summary: A large number of posts with the same content simultaneously appeared on Facebook, all of them contain this passage: “Just a little (outbreak of) pneumonia in the mainland, only a few people died (for that) in a population in 1 billion. And you’re already going crazy.”
Background information: After the first confirmed case in Taiwan reported on Jan. 21, the public began to be anxious about pneumonia, and related reports surged. At the same time, messages comparing ongoing pneumonia with influenza emphasize that the pneumonia epidemic is not as severe as flu had begun circulating.
In the afternoon of Jan. 22, a lot of posts with the same content appeared simultaneously on Facebook: “Just a little (outbreak of) pneumonia in the mainland, only six people in the population of 1 billion died (for that). And you are all going crazy. AIDS claims 770,000 deaths a year, but you guys neither put on condom nor have less casual sex.”
Authors of these posts include idol girls, live streamers and online opinions leaders (with less influence) willing to get sponsorship. The number of posts surges overnight, and the style is similar to the “#vowing_my_intention_to_vote” (“#宣告我的投票意志”) (Web Archive) campaign during presidential election⁶ which had brought the attention of information warfare on Facebook, therefore attracted attention from the public and media. These posts didn’t disseminate widely as “#vowing_my_intention_to_vote ” related posts and were deleted immediately soon after being reported.
Coherent with our previous observation of “#vowing_my_intention_to_vote ” campaign as a wave of disinformation attack last year, we suggest that this wave of “just a little (outbreak of) pneumonia in Mainland” messages is also a campaign carried by sponsorship, initiated by online marketing companies and several individual live streaming platforms. The motivation for this campaign is purely business.
Coverage: These messages were widely circulated from Jan. 21 — Jan. 23. We had archived 54 related posts. However, since Wuhan authorities officially declared stringent travel restrictions immediately after this wave of campaign had caught public attention (that is, Jan. 23,) which suggesting the epidemic became out of control in China, the dissemination of these posts dropped quickly.
Case 3: Comparison with US influenza
Summary: Many posts simultaneously appeared in different Facebook groups, emphasizing influenza, or HIV epidemics in the U.S. that caused more deaths but received less attention from the government. This campaign began at the same time with “just a little (outbreak of) pneumonia” campaign (as described in Case 2) in Taiwan.
Background information: On Jan. 16, the Weibo account “20届菜鸡考研生王一刀” published the first post, claiming that “influenza infects tens of millions in the US each year and claims nearly 5,000 deaths.” (Web Archive)⁷ Similar articles had begun to appear on Weibo. On Jan. 20, articles containing passage like “6,600 deaths by influenza in the U.S. each year” first appeared on the content farm “微杂志” (“Micro-Magazine,” now deleted, Web Archive)⁸, and were then reposted by iFeng.com (Web Archive)⁹ and www.thepaper.cn (Web Archive)¹⁰ (two influential commercial media in China).
Similar contents appear on the report of local media of Taiwan on Jan. 21. On the same day, hundreds of post containing “#美国爆发流感” (“#outbreak_of_influenza_in_US”) appeared on Weibo, and “6,600 deaths in the U.S. claimed by influenza epidemics” became the trending search on Weibo on Jan. 22.
On January 21, a large number of posts with contents like “4,800 deaths claimed by influenza epidemics in the U.S., which is more serious than those in the mainland (China), and Taiwanese government does not place (travel) restriction toward the U.S. for political concerns” appeared on political groups and pages of Facebook, many attached commentary videos made by YouTubers¹¹. On the evening of Jan. 22, senior media commentator Shaw-Kong Jaw (趙少康), a founding member of the New Party (a pro-unification political party in Taiwan), expressed a similar statement on his political show¹². Taiwan CDC refuted this kind of argument on Jan. 23, echoed by some online opinion leaders and political page on Facebook with related statistics.
Coverage: This type of disinformation was widely circulated between Jan. 21 — Jan. 23, and 29 user reports were filed to Cofacts database in 3 days¹³. (Four analyses of user-filed reports on Cofacts：1st, 2nd, 3rd and 4th) Related reports had been filed at the end of Feb¹⁴.
Phase 2: Jan. 24 — Jan. 30, 2020
The public became more anxious gradually since the lockdown of Wuhan on Jan. 23. Since it was Chinese New Year, discussions on social platforms weren’t keen, though media reports on Wuhan pneumonia were frequent, and panic buying of masks were observed. However, the severity of the ongoing epidemics had not been determined by media, since China suppressed epidemic-related information heavily and WHO assessed the global risk level to be “low” during between Jan. 23 –25, therefore opinions were suggesting “COVID-19 epidemic is less serious than what is reported.”
According to the research of Nicholas DiFonzo, rumors can help people to feel anxiety, loss of control, or a sense of needing patterns for understanding the world when feeling threatened or helpless.¹⁵ Under this view, the main purpose of the rumors currently abounding about COVID-19 is not to explain COVID-19’s explosive spread, but rather to provide remedies for its treatment and prevention.
Most of these messages had been fact-checked by Taiwan FactCheck Center or MyGoPen as false and filed to Cofact by users. User reports about these messages are still frequently filed to Cofact at the current time, indicating a wide circulation of these messages.
This phase consists primarily of two central themes: “cures” and “origins.” The six narratives of media bias, government cover-ups, COVID-19 passing a “point of no return,” seizing important information unknown by others, opposition/loathing toward China, and taking precautionary action can be found under both of these two themes.
Case 4: Folk remedies
Summary: This type of message emphasizes that taking certain foods can effectively combat pneumonia; thus people are less likely to be infected by the virus. Among all messages collected so far, this kind has the largest share. Mentioned folk remedies including:
- Drinking tea;
- Drink garlic water;
- Mouthwash by saline solution;
- Drinking warm water;
- Vegetarian diet;
- Smoking, since nicotine kills virus;
- Cleaning nostrils by soap solution;
- Dripping white sesame oil to nostrils;
- Placing onions slice around to absorb virus particles;
- Avoiding contact with rainwater (because of raindrops contain virus)
Background information: No antivirals nor vaccine had been developed, and panic of large-scale outbreak fueled the dissemination of messages about folk remedies (mostly through LINE) for disease prevention or treatment. After Jan. 30, progresses of possible treatments were reported, e.g., Remdesivir developed by Gilead Science, Inc. was administered on several confirmed patients, and the results were optimistic.
Coverage: Taiwan FactCheck Centers issued 20 reports on this type of disinformation, from Jan. 26 to Mar. 20.
Case 5: Source of this epidemic
Summary: There are two subtypes of messages: 1) the virus came from animals, and human got infected either by direct contact or consumption of wildlife animals (with images or videos of wildlife animals attached), and; 2) the virus are artificially engineered, or the epidemic is the result of biological warfare.
Background information: This first large-scale cluster infection of COVID-19, according to initial investigations by China, occurred at the Wuhan seafood market. After the disclosure of these investigations, there were news reports on the sales and consumption of wildlife animals in that market, including bats. The circulating messages related to this subtype mostly include images and videos of unrelated events, falsely claiming to be the source of this epidemic.
According to recent genealogical analyses of SARS-CoV-2¹⁶, this virus is most similar to the coronavirus sampled from horseshoe bats (Rhinolophus spp.), which suggests its possible origination. However, the existence of an intermediate animal host cannot be determined.
SARS-CoV-2 shares similar routes of transmission with SARS-CoV, and there is 79% nucleotide identity between these two species. Besides, since there is a P4 laboratory in Wuhan (National Biosafety Laboratory, Wuhan (NBL)), near the epicenter of the outbreak, therefore many messages suggest that the virus was artificially engineered by the lab, firstly appeared on Jan. 23, and have continued to evolve. The most famous examples are shown below:
- A post appeared published content farm on Feb. 14, claiming that PLA had acknowledged “the virus was artificially engineered,” with the name of one researcher Zheng-li Shih (石正丽) of the lab.
- On Feb. 21, NTDTV published a video recorded on Jan. 31, claiming that “Russian authority proved that the virus of Wuhan pneumonia is artificially engineered.”
No concrete evidence of artificial engineering had been found and accepted by the science community. For example, Kristian Andersen et al. proposed possible origins of SARS-CoV-2 entirely without human manipulation, and refuted the possibility of human manipulation¹⁷.
Coverage: Taiwan Fact Check Center issued 6 reports on this type of disinformation, from Jan. 26 to Mar. 20.
Phase 3: Jan. 31 — Feb. 25, 2020
WHO declared the disease outbreak as “Public Health Emergency of International Concern, PHEIC”on Jan. 31, 2020, since outbreaks were also observed in several countries. The declaration is the 6th PHEIC in the history WHO, which means the outbreak is spreading, thus requires immediate action. Evacuations and limitations of direct flights to China were decided by countries, news, and discussions by the online community of the outbreak surged, along with the dissemination of related disinformation.
Masks have been enlisted as controlled commodities in Taiwan since the end of Jan., and could only be purchased in convenience stores before the implementation of real-name rationing system before Feb. 6; panic among public caused excess demand of masks, and there had been related scams and disinformation online.
Large-scale outbreaks outside Taiwan have occurred since mid-Feb.; however, discussions online dropped, though related news reports were no less frequent since the outbreak was still under-controlled in Taiwan.
Analyzed content from this phase focused its content on three primary topics: facemasks, global discrimination of ethnic Chinese, and various countries’ inability to control the epidemic. Seven narratives found across these topics included: government incompetence, reports of media bias, government cover-ups, “China is the real victim,” “COVID-19 is already out of control,” seizing important information unknown to other people, and opposition/loathing toward China.
Case 6: Mask-related messages
Summary: There are two subtypes of messages: 1) Other daily commodities such as toilet papers and diapers would be out of stock; and 2) After the implementation of real-name rationing system of masks on Feb. 6¹⁹, messages were claiming the LINE chatbot “敗口罩” (“buying_mask”) powered by the government are phishing accounts.
Background information: On January 29, Taiwan CDC announced the requisition of all masks in Taiwan, and rationed them to convenience stores, so people could purchase masks in limited quantities and stockpiling became less severe. However, panic buying and limited production of masks caused excess demand. Messages containing disinformation emphasizing “since raw materials are utilized in mask production, toilet papers, diapers, and other commodities will be in short supply” lead to panic buying of commodities. The ministry of economic affairs clarified on Feb. 6 that raw materials for mask production are different than those for diapers and toilet papers thus no shortage of commodities.²⁰ The Investigation Bureau announced that the source of such disinformation had been uncovered on Feb. 9.²¹
The real-name rationing system of the mask had been implemented since Feb. 6, that all masks were sold through local pharmacies, and all people can buy two masks in 7 days with their IDs. The open-source society helped to build LINE chatbots called “敗口罩” (“buying_mask”) and “口罩哪裡買” (“where_to_buy_mask”)²², so users may interact with the chatbot and disclose their current geolocation. The chatbot displays how many stocks of masks left in nearby pharmacies. Within one day after the chatbot went online, many messages were circulated in LINE groups, claiming that these chatbots are phishing accounts. Some of them contained “[they are phishing accounts] . . . confirmed by the party office.”
Coverage: LINE is the main platform for these messages. MyGoPen (disinformation clarification website) issued two fact-check reports for these messages from Feb. 6 to Feb. 7.
Case 7: Global Discrimination of Ethnic Chinese
Summary: Beginning on 22 February, numerous videos on global anti-Chinese sentiment began appearing on Facebook and LINE. In these erroneous videos, various countries, including Italy and South Korea, were purported to be practicing severe discrimination against ethnic Chinese persons.
Background information: Beginning in mid-February, the COVID-19 epidemic began to spread abroad, causing international panic. In some countries, acts of discrimination toward persons with East Asian facial features. On 22 February, a series of videos appeared on YouTube claiming that anti-ethnic-Chinese protests had taken place in Italy and South Korea and that protestors were forecasting the collapse of China and calling Chinese people insects and viruses. Following verification, it was revealed that the events shown in the videos were not anti-Chinese demonstrations.
Because actual instances of discrimination and exclusion of East Asian-appearing persons have truly taken place in various countries, certain malefactors have been able to use concern over existing acts of discrimination to spread disinformation about attacks on Chinese people and to confuse general society.
Coverage: These videos were spread and shared through Facebook groups and Twitter. The Taiwan Fact Check Center issued two reports on this topic between March 5 and March 10.
Case 8: “Outbreak in Taiwan is already out of control.”
Summary: On February 24, many posts simultaneously appeared on Facebook, all claiming themselves being relatives or friends of some politician or medical staffs. These authors claimed that they’re informed that the outbreak of COVID-19 in Taiwan had long been out of control. These messages attracted attention since these posts were eye-catching, their compositions highly similar, and appeared across different platforms such as Facebook, Twitter, Weibo.
Background information: On February 24, the implementation of real-name rationing system of masks was nearly completed, and there were few messages emphasizing shortage of masks, the most widely circulated one, state that (punctuations adjusted): “My mom was a classmate of councilor Chia-Yu Kao (高嘉瑜) in Taipei First High School²³, and councilor had called us by phone, telling me that Wuhan pneumonia (epidemic) is very serious, and vendors in the night market of New Taipei City had discovered so many patients who went to Taipei, and the government couldn’t trace them at all. The DDP expropriated²⁴ so may masks from vendors, and DDP members can receive masks. Councilor Kao gives some masks to us. Kao herself is a very sweet and kind person.”
A suspicious fake account posted this message on a political Facebook group. It’s worth mentioning that this message wasn’t only posted but commented in many COVID-19 related discussions, and this kind of coordinated behavior is different than that observed in “#vowing_my_intention_to_vote (#宣告我的投票意志)” campaign and “#just_a_little (outbreak of)_pneumonia_in_the_Mainland (#大陸搞個小小肺炎)” (as described in Case 2). The wordings and terms of the message are suspicious to many netizens, and legislator Kao herself also clarified later.
Messages with similar compositions were disseminated either by posts in different Facebook groups or comments in discussions. These messages also appeared on Twitter, and here are some widely-circulating ones:
- “My relative sent frogs[sic]²⁵ to Pingtung arena since there are many COVID-19 patients in Pingtung and the arena became a temporary hospital.”
- “My friend had committed suicide by taking 30 sleeping pills, and he was not received by the ER of the hospital simply for not wearing a mask, which he couldn’t get one.”
- “Residents in New Taipei City has called for help, stating that there are many COVID-19 patients in their small districts²⁶, but the news was suppressed by the police.”
- Messages attached with a photo of a armored vehicle sliding off a trailer²⁷, claiming that vehicle is full of corpses of COVID-19 patients.
- Messages attached with a screenshot of a Korean disaster film Flu, claiming cremation corpses of COVID-19 patients inside of Taipei Dome by the government.
- Messages attached with a screenshot of the same Korean film, claiming there have been many deaths claimed by COVID-19 in Tainan and corpses are disposed in river.
- Falsified images of news reports titled “President Tsai Quarantined for Suspected Infection with COVID-19” and “Information About Former President Lee’s Death from COVID-19 Censored.”
These sensational messages drew attention from netizens, and the premier of Taiwan, mayor of Pingtung and mayor of Tainan, immediately refuted these claims. The Investigation Bureau issued a press release on Feb. 28, stating that related messages were composed by Weibo user in China, and further disseminated toward Facebook groups like “兩岸三地嘴砲交流（陸陸與灣灣）” (“trash talking sharing across the strait (mainlanders and Taiwanese)²⁸ ”).
Coverage: Taiwan FactCheck Center issued five reports between February 23 and February 26.
Phase 4: Feb. 25 — Mar. 31, 2020 (Two themes; nine narratives)
On 25 February, China announced a complete economic re-opening. Although China’s official statistics showed a decreasing rate of infection, the rate of returning to work remained slow. Many local governments feared a second spike in infections and did not actively implement the re-opening policy (Web Archive), placing extreme pressure on the Chinese regime.³¹ This pressure also pressured a shift of China’s strategic objectives from outward-facing messaging to inward stability maintenance.
Meanwhile, while the global epidemic continued to worsen, WHO announced on 12 March that COVID-19 could be designated a “pandemic.” On March 13, WHO Director-General Tedros Adhanom Ghebreyesus stated that Europe had become the center of the global Coronavirus pandemic, driving a renewed spike in media content.
This phase primarily consisted of two discussion themes: the origin of the virus and various countries’ inability to control its spread. Within these, a sub-theme of “fabricated fake government documents” can be said to be the continuation of the theme of the “uncontrolled” epidemic in Taiwan, though it can also be classified as belonging to the theme of “various countries’ inability to control the epidemic.” However, unlike the theme of Taiwan’s lack of control over the epidemic as discussed in the previous phase, discussion on this theme in this phase originated in Taiwanese forums before transferring to the Weibo platform. Because of this, and because the modes of transfer tend to differ between the previous phase and this one, it is observed discretely as its own theme. The disinformative content within these themes includes nine narratives: government incompetence, reports of media bias, government coverup, “China is the victim,” “COVID-19 is already out of control,” “the DPP is politicizing COVID-19”, Chinese-American relations, various countries’ epidemic prevention measures, and “America is the backstage manipulator.”
Case 9: Forged Documents
Summary: On 28 February, numerous articles suddenly appeared on Facebook using images of fake documents, to allege that the epidemic in Taiwan was out of control, or that important political figures had succumbed to the virus, or that the government had undertaken certain fabricated measures. Because these images were purported to be official documents or announcements by government agencies, they caused public panic.
Background information: On 26 February, duplicated documents allegedly from the Taiwanese Executive Yuan appeared on Facebook, claiming that individuals could use their health insurance card to obtain ten free facemasks. On 28 February, the Executive Yuan immediately appeared to clarify the false information. Aside from the modification of publicly available government documents to manufacture disinformation, others duplicated allegedly secret documents to create the effect of a leak of confidential information or fabricated official announcements from government agency webpages. This content was spread primarily through Facebook, though some specific documents were also spread on Weibo and through content farms. Among these specific documents were:
- Forged Taoyuan Government documents claiming that after mass infection in a local hospital, the city was to be locked down;
- Forged Taipei Municipal Government documents requesting high-risk patients to travel to the hospital for vaccination from Coronavirus;
- A forged Executive Yuan website page announcing “Ministry of Interior Cybersecurity Department Internal Meeting: Decides to Implement Internet Shutdown in March, Though Domestic Internet and Some Software Will Remain Open;”
- Forged National Security Bureau documents claiming censorship of information regarding former president Lee Teng-hui’s death from COVID-19;
- Forged Ministry of Health and Welfare documents claiming 1,843 samples had tested positive for COVID-19 in hospitals in Tainan City.
The first fabricated item from the above list, “Taoyuan Municipal Government documents” could not be retrieved from where they first appeared, but images of the fabricated documents can be seen on the Weibo account of blogger “Shenzhen Xiaotian,” posted on February 29, the day that the fabricated documents were released. Additionally, the fabricated National Security Bureau documents can be seen to have been forwarded by Weibo blogger “Heichaojun” on February 29.
Because the false news content of these fabricated items could easily incite public panic, related government agencies responded to them quickly. On February 28, the Administrative Yuan made public clarification regarding the false document that, aside from the false document claiming that “Health Insurance Cards can be Used to Obtain Ten Free Facemasks.” On February 29, Taoyuan mayor Zheng Wen-Can is publicly declaring the falsehood of document alleging the closure of the city. On March 1, the Criminal Investigation Bureau issued an article to notify the public of the presence of fabricated government documents and to announce their search for sources of disinformation.
We can see these disinformative items as a continuation of the previous phase of the narrative theme that the “outbreak in Taiwan is out of control.” However, the particular dissemination methods used for “fabricated government documents” differ from those of the previous phase, notable in particular for the spread of these documents on Weibo. These distinctions mark this phenomenon worth of particular observation.
Coverage: Taiwan FactCheck Center issued 2 reports between February 29 and March 2.
Case 10: “The source of the virus is not China.”
Summary: This narrative first appeared on February 21, when posts began circulating claiming that the COVID-19 virus did not originate from China, but rather from the United States or other another country. The primary content of this narrative trend included the following sub-themes: “COVID-19 was carried into the Wuhan Military World Games by US soldiers;” “COVID-19 is just the flu [‘U.S. Influenza ’], the virus’ source is the United States;” “Experts claim that the epidemic may have already been active in Italy last November;” and related concepts.
Background information: As early as January 23, messages began spreading on Weibo alleging that “an epidemic originating from US soldiers was transmitted to China during the 2019 Wuhan Military World Games.” However, the accounts discussing such content were not those of popular Weibo bloggers, and many of these posts received short “Doubtful” responses expressing skepticism (these posts and their responses can no longer be seen on Weibo). On January 30, the public Weixin account “Xuekawu” published a post asking, “Why does Wuhan have to rely on the military for this epidemic?” (post already deleted; backup copy available here), which incited extensive discussion of COVID-19 as a US-manufactured virus, said to have been spread as an attack against Wuhan during the Wuhan World Military Games.
At 8 pm on February 21, Japanese news network Asahi published the article “Is the virus thought to be influenza actually Coronavirus? Obstacles in US medical treatment.” At 11 pm that evening, the first disinformative items began appearing on Weibo. Weibo blogger “Sima Pingbang” published the first disinformation items (Web Archive), claiming “This Coronavirus truly originated in the United States, and the United States is treating it like influenza. In that case, the Wuhan Military World Games was probably where the epidemic started, so of course it’s impossible to find a ground zero in China. We should focus this search on the US athletes.”
From this point, the narrative theme “Before the outbreak in Wuhan, COVID-19 was already transmitting all over the United States; it’s just that they were treating it as influenza” began spreading widely. On February 23, YouTube user “phoenixbinbin 安好TV“ published a video titled “Japanese Asahi news: Was the United States already experiencing a Coronavirus outbreak before Wuhan?”. This video was shared extensively on Facebook.
On the evening of February 27, on the Dongsen Zhenglun program “This! Is not news,” the New Party city councilor Pan Huaizong quoted an article from ChinaXiv, the Chinese Academy of Sciences technical discussion and publishing platform, claiming that COVID-19’s origin was the United States. On the evening of February 28, Weibo blogger “Wuxin Jianying” shared this video clip to Weibo, after which the video was distributed widely. Soon after, the video was further disseminated and shared in Taiwanese Facebook and LINE groups.
On March 2, posts began spreading on the internet stating “United States CDC confirms first Coronavirus case in the United States! China is no longer the scapegoat,” accompanied by a photo of United States Secretary of State Michael R. Pompeo. On March 5, a video began circulating on Youtube claiming “Trump finally admits: Coronavirus is just the flu” (video originally, but no longer, available here). After this, many pro-KMT Facebook and LINE groups spread immediately began spreading the video. On March 12, another video began spreading on Facebook, claiming “the United States finally admits Corona-related viruses are just the flu.” These videos were also shared to some pro-KMT groups.
In addition to the United States, Italy was also discussed as a potential source of the COVID-19 epidemic. On February 22, in an interview with the US National Public Radio (NPR), director of the Italian Mario Negri Institute for Pharmacological Research Giuseppe Remuzzi explained that a number of pneumonia cases of unknown cause had appeared in northern Italy in November and December of 2019. He hinted that China may have covered up the status of its epidemic. But Chinese state media networks CCTV, CGTN, and others immediately reported that “Expert claims the epidemic may have already been spreading in Italy by November 2019.” These reports suggested that Italy could have instead been the outbreak point of COVID-19. This reporting also drew copious discussions on Weibo.
Coverage: Taiwan FactCheck Center issued ten verification reports between March 3 and March 20.
Case 11: Various Countries’ Inability to Control the Epidemic
Summary: Beginning on the evening of March 15, a series of posts began appearing on Twitter whose content all related to to the claim that “The virus in various countries is already out of control; I have learned from family and friends that countless people are going to hospitals but hospitals have no means to test for the virus and can only dismiss patients and send them home…I have already booked my tickets back to China, by focusing our efforts right when it matters, we can make a big impact.” Related fake news items also began appearing on LINE and Facebook, claiming that European and American countries were facing large numbers of casualties and societal breakdown because of the epidemic.
Background information: On March 12, WHO announced that COVID-19 could be defined as a “pandemic.” On March 13, WHO Director-General Tedros Adhanom Ghebreyesus said that Europe had become the global center of the Coronavirus pandemic.
On the evening of March 15, a large quantity of duplicated news content suddenly began appearing on Twitter. The content of these posts stated “The virus is already out of control in several countries. I have heard from friends in hospitals in several countries that countless people are coming to the hospitals every day for diagnosis, but the hospitals have no test kits, so all they can do is send the people back home. Several countries have many elderly people, and many citizens of these countries are dying in their homes. Because they can’t confirm the diagnosis, they don’t count them as infected, and that’s why the countries maintain such a low [virus] growth rate…it’s really scary. I have already booked my flight back to China, if we focus our efforts when it matters, we can do big things.”
Because many of these posts were duplicated news items, internet users quickly perceived that it was a disinformation attack. On March 16, many Twitter users began using the same formulation to spoof these posts, spreading content like “the virus is already out of control in the Milky Way.” On March 17, the media also reported on these disinformation efforts(Web Archive).³²
It is worth noting that on the afternoon of March 18, Ifeng Finance published an article (Web Archive) combing through Indian, Russian, and other foreign resident Chinese Weixin public accounts, publishing an alternate version of the narrative from various countries. All of these posts followed an identical format: “Epidemic in [country]: stores have all closed up shop; it’s too hard for Chinese to do business!”; only the name of the country changed from message to message. The information contained in these messages originates from “independent media marketing accounts”³³ and are likely posted in return for financial gain. That evening, Shanghai United Media Group-owned ThePaper.cn published an article (Web Archive) claiming that all of the related posts originated from the work of a single actor under the alias “Guohong.”
On March 19, several Weixin public accounts, including “Chuangshiji” (Web Archive), “Matou Qingnian” (Web Archive), and others, began referencing the “out of control” tweets in their posts. We believe that these were also the work of Guohong.
At noon on March 20, tech.sina.com.cn (“Xinlang Keji”) issued a press release stating “‘Weixin’ public accounts recently publishing content using the “Epidemic in [country]: stores have closed up shop, it’s too hard for Chinese to do business!” text are in violation of the “Weixin Public Platform Terms of Use”…we have deleted one hundred posts and blocked 50 accounts and three groups in violation [of Weixin policies].”
In addition, at the end of March, videos began appearing on Facebook about various countries’ loss of control of the epidemic, featuring the following content:
- The military has entered the city of New York. The United States government may fear a repeat of the 1977 New York Blackout and the resultant robbery and violence. Today, the military has already begun establishing itself in the city. In a familiar sight for ethnic Chinese, today, we have already heard that the Chinese of New York’s Chinatown is being mugged, and the glass has been broken on storefronts of Chinese businesses and banks in New York’s Flushing neighborhood.
- A supermarket in Louisiana has been overrun. No one manages the tills, and vandals take what they want. Soldiers holding guns cannot control the situation; they are only there to prevent conflict and bloodshed. Customers come and go as they please.
- The elderly face the highest risk of contracting COVID-19. Italians have quarantined persons older than 65: a cruel situation.
Coverage: Taiwan FactCheck Center issued 3 verification reports between March 27 and April 1. MyGoPen released 2 reports in that same period.
(2) Transition between Different Patterns of Disinformation Dissemination
Following its analysis of COVID-19 related disinformation, Doublethink Lab conducted a comparative study between the dissemination methods of these items and those of disinformation activities in 2018. Based on this comparison, we have discovered that dissemination strategies have been continually changing. This study allows us a glimpse behind the scenes of dissemination strategy and transformation.
Analysis of Dissemination Modalities
The table below uses as its basis of analysis the four modes proposed in Dr. Puma Shen’s “Exploration of China’s Cognitive Domain Operation Modes: The case of Taiwan’s 2020 elections.”³⁴ This methodology allows for a further dissection of the methods by which COVID-19 related disinformation has been spread.
The best-known disinformation campaign involving Taiwan in 2018 could be none other than the Guanxi Airport Incident. Investigating the dissemination channels of this event reveals that related disinformation content first appeared on the Weibo page of Chinese local television news reporter “Hongshui Mengshou baby,” who made a post mentioning Communist Youth League. This post was deleted two days later, but not before the content of the post was shared to the “Guanchazhewang” (“Observernet”) content farm.(Web Archive)³⁵ By the following day, the post had been shared to Taiwan’s PTT forums and Facebook fan pages, drawing considerable public attention and then becoming a central topic of political talk shows and a frontpage news story.(Web Archive)³⁶ Through what is suspected to be an order from the Communist Youth League (the Communist Youth League has admitted in other cases to connections with pro-China “cyber troops,” often known in Chinese as “小粉紅,” “little pinks”), this mode most resembles the “Outreach Mode.” If the Communist Youth League were not involved, then it would instead be classified as a general “Pink Mode” operation.
In the 2020 COVID-19 related disinformation campaign, the disinformation dissemination channels of articles related to “Comparisons between pneumonia and U.S. Influenza ” are similar to those of the Guanxi Airport incident. In both cases, information was spread from Weibo pages to content farms and then into news items in Chinese media, and finally into Taiwanese media through the “Outreach Mode.”
At the same time, the copiously distributed, suspected disinformation narrative of “just a little (outbreak of) pneumonia in the mainland” was disseminated primarily through posts on individuals’ Facebook pages. After investigation, it was discovered that this message was spread by internet celebrities under the same management company, thereby more resembling “Collaboration Mode.” #vowing_my_intention_to_vote ” disinformation campaign that took place prior to last year’s elections also employed the “Collaboration Mode,” but while Instagram accounts were also targeted in the “#vowing_my_intention_to_vote ” disinformation campaign, Instagram was not targeted during the 2020 “just a little pneumonia” campaign. One possible reason for this may be that most Taiwan internet marketing companies are still defining their operation strategies on Instagram, whereas the businesses operating on Facebook platforms are already very familiar with the website’s operations.
Furthermore, the volume of embedded marketing accounts found to be involved in the “just a little pneumonia” campaign fell far short of those in the “#vowing_my_intention_to_vote ” campaign, and the number of accounts was far smaller, as well. Furthermore, this campaign was detected as a disinformation attack almost immediately after it was launched. Because of this high detectability, we assess that “template duplication” attacks will become less and less popular. Moving forward, we anticipate that Collaboration Mode operations will instead find more effective and sophisticated attack strategies.
Disinformation related to “the virus is out of control in Taiwan,” aside from posts on individuals’ Facebook walls, the majority of messages and posts are also made in Facebook groups, and we have not seen a large number of posts appearing on Facebook fan pages. This type of disinformation campaign usually involves fake Facebook accounts. However, because there is no clear evidence at present that marketing companies have used online “influencers” as agents for their operations in this campaign, this activity would more appropriately be categorized under “Pink Mode” operations. Regarding “various countries’ inability to control the epidemic” messaging, messages have only appeared on Twitter, and the related accounts are all fake or individual accounts. None are internet celebrities contracted by marketing companies, so this campaign should also be considered a “Pink Mode” operation.
Regarding the different attack modes, China has primarily utilized “Outreach Mode” and “Pink Mode” toward Taiwan. During the Taiwanese election, “Content Farm Mode” was more commonly seen, because of Facebook’s controls forcing similar attacks to shift to YouTube channels. In the “Collaboration Mode,” because the epidemic situation in Taiwan was relatively stable, the general “collaborator” could not find a point of attack, which granted Taiwan some breathing room. However, as soon as an attack point appears, we believe that “Collaboration” and “Content Farm Mode” operations could be immediately activated. Because several content farms have been established since March, producing primarily content related to Coronavirus and the popular campaign to recall Kaohsiung City Mayor Han Guoyu (韓國瑜). These content farms will continue to require close attention.
The Transformation of Communication Media
Based on the discoveries of our past observation, some content farms and Facebook fan groups employ a particular business model, in which content farms open backend permissions allowing administrators of fan pages to post articles on their pages in cooperation with these content farms. Later, content farm managers will tally how many posts content propagators have made on their respective pages and share a portion of the economic proceeds with those propagating agents.
This business model had suffered since 2019 when Facebook launched a comprehensive takedown of specific content farm links. After that, certain content farm links could no longer be posted because they “violated [Facebook’s] community guidelines.” Although content farms have tried to employ multiple web domains to evade Facebook’s restrictions, they have not shown any clear success. Indeed, among the 2020 COVID-19 disinformation content, relatively few content farm articles appeared, and they were transmitted only through LINE groups.
In place of content farms, many YouTubers are now making commentary videos. These “political YouTubers” began appearing in large numbers in 2019 and distributing their videos through LINE and Facebook groups. It is worth noting that these political YouTubers will all release videos on the same topic in the same narrow timeframe in order to shape public opinion. In fact, these videos are often even release before these topics can reach Taiwan’s political talk shows.
For example, during the 2020 wave of false information around COVID-19, simultaneous with the “just a little pneumonia in the Mainland” attack, YouTubers “Hanguoren,” ”Xiangdian Jiudian,” “SHA SA Sasha Sasha,” and ”LTM News [Zhongdian Zixun],” were all posting related videos to Facebook.³⁷ All of these videos highlighted that the death count from US influenza was higher than that of COVID-19. Just one day later, Shaw-Kong Jaw (趙少康) was discussing these same points on his political talk show.
There are other important characteristics of videos and channels promoting related disinformation content. For example, channel subscribers, likes, and comments are scant, but view counts are very high. Research suggests that dissemination methods usually involve sharing YouTube videos to LINE groups, where recipients can watch these videos directly on the LINE platform. Recipients generally will not be directed to the YouTube platform. However, their views are still counted by the YouTube algorithm, producing an observable anomaly where videos display high view counts but very few comments.
Aside from evading the standards of Facebook’s community guidelines, this transition of disinformation transmission methods also allows attackers to take advantage of YouTube’s recommendation algorithms. These algorithms direct viewers of a video to more videos with similar content or other videos on the same channel, thus helping the manufacturers of the disinformation the optimize user adhesion. Moreover, using videos to transmit disinformation can effectively sidestep third-party fact-checking activities in LINE, creating difficulties for detection and reporting.
It was previously believed that the primary battlefield for influencing public opinion was the PTT forum platform, but in this new wave of disinformation attacks, it seems that PTT forums are playing no significant role. LINE groups, on the other hand, remain a primary staging ground for disinformation operations. Additionally, disinformative content relating to the themes of “the epidemic in Taiwan is out of control” and “various countries’ inability to control the epidemic” have also appeared extensively on Twitter. It is worth continued investigation whether Twitter’s role in this campaign signifies that the platform is becoming a new important channel for disinformation transmission.
Separation of Attackers and Distributors
Compared to the information spread of 2018’s Guanxi Airport incident, the manufacturers and distributors of 2020 COVID-19 disinformation are relatively disconnected.
Disinformation from the 2018 Guanxi Airport incident can be traced back to the Communist Youth League, whence it was spread through political Weibo accounts as agents. After the creation of a large number of fake news items, a network of observers worked to turn disinformation content into news headlines. Business and media entities then got involved, and internet commenters then put news content into PTT forums, whence they could directly influence public opinion after this content was being found and incorporated into Taiwan’s political talk shows, where they could directly influence public opinion. The interesting thing is that this mode does not involve intermediary collaborating agents in the distribution process.
Similarities do appear between information dissemination channels of the 2020 COVID-19 disinformation theme of comparing Coronavirus to U.S. Influenza and those of the Guanxi Airport incident. In both cases, the content was shared from Weibo pages to content farms. From there, Chinese media picked up the story in their news reports, and those Chinese reports were then covered in Taiwanese media. However, it has been impossible thus far to confirm an orchestrator of this disinformation stream.
Both the “just a little pneumonia in the Mainland” and 2019’s “#vowing_my_intention_to_vote ” disinformation campaigns employed “Collaboration Mode” dissemination. In these instances, once the veracity of posted content was called into question, the posters themselves quickly deleted it, sometimes even publishing posts explaining their actions. From this behavior, we can observe that disseminators were not acting with a political motivation, but rather were distributing content as simple intermediaries on behalf of purely economically motivated rent-seekers. These purely economically motivated distributors may themselves have no knowledge of who manufactured the disinformative content, and they often do not know the political motivations behind the content’s spread. This fissure between creators and distributors complicates disinformation tracing.
“Collaboration Mode” refers to the distribution of false information through local agents, meaning that disinformation manufacture and dissemination are both performed by Taiwan-based actors. Although it is possible to use social network monitoring tools to trace the accounts used for disinformation dissemination, what motives are driving these distributors? What is their motive to cooperate with attackers (previously suspected creators of disinformation content)? It is difficult to discern whether these motives are economic or political. And because this disinformation content is manufactured by Taiwanese collaborators, the potential of this content to create influence cannot be understated.
Additionally, the “Pink Mode” disinformation of the “Taiwan’s epidemic is out of control” narrative demonstrates that disinformation content no longer originates from a single unified manufacture node, but rather is generated by disseminators themselves. The decentralized nature of this mode of dissemination makes it harder to detect, suggesting that this mode may become more prevalent in the future.
External Propaganda as a Domestic Stabilization Measure
Based on analysis of the disinformation narrative “the source of the virus is not China,” we can discover that the disinformation content received in Taiwan was segmented and arrived long after its appearance in China. For example, the false headline “Japanese Asahi Network Says Coronavirus is Just U.S. Influenza a” only reached Taiwan as spillover from Chinese Weibo dissemination. We can conclude that this story’s primary objective was to support Chinese “stabilization efforts” (維穩). In contrast, the narrative that “Coronavirus is just U.S. Influenza ” was immensely popular on Weibo, leading to its spread to Taiwan, where it drew the attention of the Taiwanese public.
Further examination of Chinese social platforms and news shows this narrative’s continuing ebb and flow:
On March 12, the Chinese Minister of Foreign Affairs Zhao Lijian generated large controversy after tweeting, “the virus might have been brought to Wuhan by the American military.” (“病毒可能由美軍帶來武漢”)³⁸ A few hours later, Ministry of Foreign Affairs spokesperson Hua Chunying tweeted, “the 2019 U.S. Influenza epidemic was actually Coronavirus, so it is inappropriate to call COVID-19 ‘Chinavirus.’” These Tweets drew a wave of discussion on Weibo, where it was commonly assumed that the Ministry’s actions were intended as a counter-attack against United States discrimination against China.
However, merely four days later, on March 16 Chinese state media outlet China Daily posted a tweet on its official account, clarifying, “The five foreign competitors in the Military World Games were hospitalized for treatment of malaria; no connection to COVID-19” (no longer available, see image below). It is believed that this tweet represents a comprehensive negation of Foreign Minister Zhao Lijian and Spokesperson Hua Chunying’s tweets. According to Reuters³⁹, the Chinese Central Daily News tweet may have been a result of the US disapproval of the Foreign Ministry’s statements, communicated during the US Secretary of State’s March 13 meeting with China’s US ambassador. Other media agencies have also speculated that US Secretary of State Mike Pompeo may also have expressed strong opposition to China’s blame-shifting over the “Wuhan epidemic” to the United States during his March 16 call to CCP Politburo member Yang Jiechi (Web Archive)⁴⁰. However, because China Daily chose to use traditional Chinese characters in its tweet, their clarification on Twitter did not receive a significant response on Weibo.
On March 22, the Chinese embassy to the United States released on its official website the full text of a March 17 interview between Axios, HBO, and Chinese ambassador to the US Cui Tiankai. (Web Archive)⁴¹ The text includes questions about the source of COVID-19, to which Cui responded, “The rumor that Coronavirus originated in an American military lab is crazy.” (“散播新冠狀疾病源頭來自美國一個軍事實驗室的謠言，是『瘋癲』的“)⁴² He further clarified that Zhao Lijian’s tweets do not represent the Chinese Communist Party’s views. After this text was released, it was briefly circulated widely through public WeChat accounts, but there was no significant discussion thereof on Weibo.
In order to understand the changes in China’s internal public opinion, we must shift our tracing efforts to Chinese media. Doublethink Lab’s Chinese Media Database monitors 1,423 Chinese news websites, including Weibo, WeChat and other predominant Chinese social platforms. This database stores 30–50,000 articles and posts every day, reaching almost five million items as of the end of March 2020.
Based on the materials stored in the Chinese Media Database, we discovered that the circulation of posts related to “US Military World Games” began in February 2020 but declined substantially after March 22. Between March 16 and March 23, Chinese officials set the tone for discussion of content like “COVID-19 was brought to the Wuhan Military World Games by an American soldier,” beginning a national silencing of such content.
In late March, Taipei City Councilor Pan Huai-Zong told Taiwanese EBC News “This! It isn’t news” regarding political talk show discussion of the United States as a source of COVID-19. His comments were later subject to another round of “stabilization and peacekeeping” by Chinese authorities after they were shared on Weibo. This Chinese government response to a Taiwanese soundbite has never been seen before, and the underlying motivations for this response merit deeper study.
Furthermore, in the case of the “various countries’ inability to control the epidemic” theme, we can observe that the posts that spread across Twitter were a spillover from an internal Chinese censorship crackdown but afterward were quickly addressed by Chinese state media and other platforms. This was ultimately reported on in the media. (Web Archive)⁴³ Whether China’s shutdown of this content stemmed from domestic concern that China’s medical facilities could not support the medical needs of Chinese students returning from study abroad, or from some other reason, also deserves further observation.
4. Conclusion and Limitations of This Study
From our analysis of the 2020 disinformation related to COVID-19, it is clear that not every controversial news item broke out because of a connection to China. However, it is certain that many of these controversial or false news items were widely disseminated precisely in sync with Chinese propaganda narratives.
We’d observed the evolution of disinformation campaigns and tactics of information operation. The main paths of dissemination have change from legacy media, online media or public Facebook pages that can be monitored easily (with the help of crawlers), to Facebook groups, personal profiles or closed communication services (i.e., LINE, WeChat, etc.) that are harder to be accessed and archived automatically by crawlers. Contents of dissemination have shifted from text to live streams and videos, which are harder to be collected and monitored. Direct sharings became less frequent, and the fragmentation between content manufacturers and disseminators has been more common, which is driven by purely business motives. Such fragmentation would make the dissemination of disinformation more private, and targeting to smaller/more specific audiences.
Disinformation related to people’s livelihood or political issues are tended to be transmitted more. Even when official or unofficial verification bodies clarify the facts of these stories, they continue to be widely shared on the internet, particularly through private LINE and Facebook groups. The actual influence of this content is broader than what can be publicly observed, and it is relatively difficult to analyze the sources or motives behind their spread.
However, it is worth noting that as general society continues raising their vigilance toward controversial news, what impact will these “failed disinformation attacks” have on society? During the 2020 COVID-19 disinformation campaign, clearly unsuccessful narratives like “just a little pneumonia in the Mainland” and “Taiwan’s epidemic is out of control” were all clarified and covered in mainstream media within a few days of their appearance.
The “#just_a_little (outbreak of)_pneumonia_in_the_Mainland ” was clearly implemented by rent-seekers for economic motives. The failed “#vowing_my_intention_to_vote ” attack once again used remarkably similar attack methods and quickly failed. Because these similar methods had already failed once, why were they used again in a later campaign? Furthermore, during the “Taiwan’s epidemic is out of control” attack, because the content of the disinformation items was so exaggerated, using words uncommon to Taiwanese Mandarin speakers, this campaign was also immediately detected. Were the attackers behind these quickly-discovered truly so ignorant of Taiwanese culture? After all, perhaps their only goal was to create confusion. Or did they not care whether or not their attack succeeded?
Another possibility remains that the agent/intermediary nodes of these campaigns were simply task-oriented or budget-focused implementers, but this sort of possibility requires further investigation and verification.
Otherwise, as previously mentioned, although it is difficult to confirm the instigators of many of these attacks, the timing of the attacks adheres closely to related events relevant to China’s internal politics and public opinion. We can begin to determine thereby that much disinformation dissemination is not external propaganda from China but rather is intended as an internal measure of stabilizing control on Chinese society. The interplay between China’s internal controls on public opinion, the narratives of its external messaging, and how China uses both to promote stability at home, still requires deeper study.
Research Limitations
This study was restricted to material originating in the period from January 1 to March 31, 2020. As COVID-19 continues its global spread, we continue observing the dissemination methods for COVID-19 related disinformation. The materials used for the preceding analysis were collected by Doublethink Lab, who stored suspect Facebook fan pages, Facebook public groups, and Chinese web sources of news items, including state media pages, Weibo, and public WeChat accounts. This study also references a cross-section of information from the Taiwan FactCheck Center, public reports released by MyGoPen, and the Cofacts database.
Reference
[1] Brendan Nyhan and Jason Reifler (2010). When Corrections Fail: The Persistence of Political Misperceptions, Political Behavior 32, 303–330. DOI: https://doi.org/10.1007/s11109-010-9112-2
[2] The first officially confirmed case in Taiwan was reported in Jan. 21, 2020, who arrived Taiwan from Wuhan, China in Jan. 20, 2020.
[3] “嚴重特殊傳染性肺炎”, which is the official naming of COVID-19 in Taiwan, long before the naming decision by WHO in Feb. 11, 2020.
[4] The announcement of Taiwanese CDC is made public in Jan. 10, see https://www.cdc.gov.tw/Disease/SubIndex/N6XvFa1YP9CXYdB0kNSA9A (in Traditional Chinese, last accessed Mar. 29, 2020)
[5] Two analyses of user-filed reports on Cofacts: the first one, the second one.
[6] Local news report on 2019.12.20 (in Traditional Chinese): https://newtalk.tw/news/view/2019-12-20/343092 (Web Archive)
[7] http://weibo.com/5519966859/IpO7y8blS (now deleted)
[8] Cache: https://zazhi.qunba.haiwai.com/m/wenzhang/1602050 (Web Archive)
[9] http://news.ifeng.com/c/7tOiJKAdDPo (Web Archive)
[10] https://www.thepaper.cn/newsDetail_forward_5580207 (Web Archive)
[11] E.g. https://www.youtube.com/watch?v=21JFf6uVa_w (Last accessed May 4, 2020.)
[12] See https://www.youtube.com/watch?v=PX0kjSaGJVQ (related media report in Traditional Chinese, last accessed Mar 30, 2020.)
[13] Four analyses of user-filed reports on Cofacts：1st, 2nd, 3rd and 4th.
[14] The analyses of user-filed reports on Cofacts until the end of Feb.
[15] Nicholas DiFonzo, The Watercooler Effect: An Indispensable Guide to Understanding and Harnessing the Power of Rumors, Avery, 2008
[16] Report of the WHO-China Joint Mission on Coronavirus Disease 2019 (COVID-19) https://www.who.int/docs/default-source/coronaviruse/who-china-joint-mission-on-covid-19-final-report.pdf
[17] Kristian G. Andersen, Andrew Rambaut, W. Ian Lipkin, Edward C. Holmes & Robert F. Garry (2020). The Proximal Origin of SARS-CoV-2, Nature Medicine. https://doi.org/10.1038/s41591-020-0820-9 (Online article, last accessed Mar. 30, 2020)
[18] See the announcement by WHO: http://www.euro.who.int/en/health-topics/health-emergencies/international-health-regulations/news/news/2020/2/2019-ncov-outbreak-is-an-emergency-of-international-concern (last accessed Mar. 30, 2020)
[19] Local news report (in English) on the implementation of real-name rationing system: https://www.taiwannews.com.tw/en/news/3872149 (last accessed Mar. 31, 2020)
[20] See the announcement by the Ministry of Economic Affairs on its facebook page (in Traditional Chinese): https://www.facebook.com/moea.gov.tw/photos/a.335602533146596/3657184090988407/?type=3 (last accessed Mar. 31, 2020)
[21] See the local news report (in Traditional Chinese): https://news.ltn.com.tw/news/society/breakingnews/3062645 (last accessed Mar. 31, 2020)
[22] These chatbots are supported by open source society in Taiwan, and powered by mask data API controlled by the Bureau of National Health Insurance. For details of this chatbot, see this local news report (in English): https://focustaiwan.tw/society/202002280019 (last accessed Mar. 31, 2020)
[23] Kao was a former councilor of Taipei, and is currently a member of current ruling party DDP. She was elected legislator of Taiwan in the January election. She is an alumnus of Taipei First Girls’ High School, and so called “Taipei First High School” is the archaic name of another all-boy school (now Jianguo High School) used before WWII.
[24] The original term in this message is “查收”, which precise meaning is “check and receive” in Taiwanese Mandarin, and is mostly used when receiving packages or registered mail. According to the context of this message, the term is suggesting requisition of private properties followed by unfair/undue distribution (“to DDP member”) and “expropriated” or “confiscated” may be adequate translations.The imprecise usage of “查收” is another evidence showing that author of this message is not familiar with the nuance of Taiwanese Mandarin, hence unlikely to be a Taiwanese him/herself.
[25] Although frogs (more specifically: Rana catesbeianus or Hoplobatrachus rugulosus) could be foodstuff in Taiwanese cuisines, they’re not common in daily diet and unlikely to be included in meals serving patients.
[26] The original term in this message is “小區”, i.e. “small district” which is the basic unit of urban community development and regulation in China. This term is commonly used in China but rarely used in Taiwan.
[27] The photo was taken as a press photo about an military accident and totally unrelated to COVID-19 outbreak.
[28] The literal translation would be: “trash talking sharing for two shores (“兩岸”, i.e. Taiwan and mainland China) and three territories (”三地”, i.e. Taiwan, mainland China and Hong Kong) (mainlanders and Taiwanese)”. Hong Kong (and Hong Konger users) is actually omitted in the title. Also, the term “灣灣/湾湾” specifying “Taiwanese” (and “陸陸/陆陆” is the reciprocal, specifying “mainlanders”), is a term which can express taunt or intimacy, depending on the context. For example, a newscaster of CCTV saying “湾湾, 回家吧” (“Come back home, Taiwanese”) would conceived as frivolous, even contemptuous, by certain Taiwanese for the undue intimacy suggested the term. “湾湾” is commonly used by (mainland) Chinese netizens to denote Taiwanese (and “陆陆” is used reciprocally to denote themselves), but Taiwanese netizens rarely used “陸陸” and “灣灣” to denote both of them. Therefore, it’s an evidence indicating that administrators of this group are either (mainland) Chinese netizens, or less likely, some Taiwanese netizens who is familiar with online culture in China and politically/culturally pro-China.
[29] From https://twitter.com/YHzGwpQzpm7zomh/status/1232372040927080449 (now deleted).
[30] From https://mobile.twitter.com/RF3body_chinese (now deleted)
[31] “Dongguan re-opening rate at 2%, Ningbo almost 40%! Xi Jinping calls for re-opening but instructs not to compel local governments; what’s really going on behind the scenes?” (https://www.businessweekly.com.tw/international/blog/3001839）(Web Archive)
[32] “Fifty Cent rumors that various countries have lost control of the viral epidemic, posting completely duplicated text; Taiwanese netizens poke fun.” Source: https://newtalk.tw/news/view/2020-03-17/376212）(Web Archive)
[33] “Weibo Marketing is a marketing service for business owners and other individuals who create value or conduct operations on the Weibo platform. It is also a means for business owners or individuals to satisfy customer needs. Weibo Marketing uses Weibo as a marketing platform, every user (fan) is a potential marketing target. Entrepreneurs can update their microblog to broadcast business and product information to internet users or establish a new image for their business or products. By updating content every day, business owners can communicate and interact with everyone or publish the content most interesting to their followers. Reaching marketing objectives in this way is the method promoted by Weibo Marketing.” Li Jun, “Big Data” (Chinese title: “實戰大數據 — 客戶定位和精準營銷”), Songbo Publishing, February 1, 2019.
[34] Shen, Puma, “中國認知領域作戰模型：以2020台灣選舉為例 (Exploration of China’s Cognitive Domain Operation Modes: The case of Taiwan’s 2020 elections)” (Working Paper).
[35] “We’ve reached this point without realizing it, Chinese consul receives a Taiwanese compatriot who asks…” Source: https://www.guancha.cn/internation/2018_09_05_470949.shtml (Web Archive)
[36] “Dr. Puma Shen Publishes: ‘The True Story Behind the Guanxi Airport Incident,’ Criticizes [Both Taiwanese Political Parties] for Not Verifying Disinformation.” December 02, 2019. Newtalk.tw, Source: https://newtalk.tw/news/view/2019-12-02/334745 (Web Archive)
[37] Respectively, “寒國人”, “想点就点,” “SHA SA 萨莎 Саша,” “LTM News【重點資訊】”
[38] “Original: “病毒可能由美軍帶來武漢”
[39] https://www.reuters.com/article/us-health-coronavirus-china-diplomacy-idUSKBN2102XW
[40] https://www.rfa.org/cantonese/news/us-phone-03172020063039.html (Web Archive 版本)
[41] http://www.china-embassy.org/chn/sgzhichuang/dshdong/t1759545.htm (Web Archive 版本)
[42] Original: “散播新冠狀疾病源頭來自美國一個軍事實驗室的謠言，是『瘋癲』的“
[43] “Experts: Chinese Announces Opposition to Extensive Misinformation about Coronavirus; Trump’s Counterattack Proves Effective.” Source: https://www.voachinese.com/a/experts-chinese-fake-news-on-coronavirus-hits-overseas-chinese-trump-push-back-beijing-narrative-20200320/5338755.html (Web Archive)
