On January 11, 2020, Taiwan held its 15th presidential and 10th Legislative Yuan election. Taiwanese citizens soundly re-elected Democratic Progressive Party (DPP) candidate, Tsai Ing-wen, who won 57.1% of the vote over her opponents, Kuomintang (KMT) candidate Han Kuo-yu (who took 38.61%), and the People’s First Party candidate James Soong (4.26%). The DPP also maintained its majority in the Legislative Yuan, though with a slight decrease of a few seats. Voter turnout was high, with almost 74% of eligible voters casting ballots, up from 66% in 2016. According to Liberty Times Net, the number of Taiwanese voters living overseas who registered to vote on Saturday more than doubled from 2016.

In this blog post we provide an overview of some unexpected results from the 2020 election. We examine how the election was perceived on social media, within each candidate’s respective Facebook fan Pages and Groups. We wrap up this election coverage series by looking at areas that are important battlegrounds in Taiwan for ongoing disinformation dissemination, such as YouTube, and on the role the perceived threat of foreign interference played in this election. Lastly, we end by focusing on why collaboration to detect and mitigate influence operations is so important.

Presidential and Legislative Yuan Election Surprises

This election set a few historical precedents both for Tsai and for the presidential election process. According to the New York Times, Tsai received more votes and a wider margin of victory than in 2016. Tsai also received more votes than “ever recorded for any candidate in a presidential election in Taiwan” as reported by Focus Taiwan, the English section of the Central News Agency. This result was a dramatic turnaround given her approval ratings at the beginning of 2019. Experts assessed the Hong Kong protests to have been a significant factor – if not the largest factor – in helping Tsai regain her support. 

Out of 113 seats, the DPP maintained its majority but dropped from 68 to 61 seats. The KMT gained 3 seats to increase its total to 38. Taiwan’s People Party (TPP), the party of Taipei’s mayor, received a good portion of the votes, winning 5 seats but just 1.5 million votes compared to 4.8 million and 4.7 million won by the DPP and KMT respectively. The remaining 9 seats went to various other parties.  

Some results in the Legislative Yuan races came as a surprise. One headline from Focus Taiwan read, “Young candidates, underdogs prevail in several legislative races”. One contested race was between Bo-Wei Chen (陳柏惟) of the Taiwan Statebuilding Party, and incumbent Yen Kuan-heng (顏寬恆) of the KMT. Chen, who was supported by the DPP, defeated Yen in a close race claiming 51.15% of the vote to Yen’s 48.85%. This result was something of a shock, since the KMT candidate came from the Yen family, which has governed the district for more than a decade and is described as a “local political dynasty” by Focus Taiwan. Another surprise victory was Saidai Tarovecahe (伍麗華), who became the first member of the DPP to represent the Highland Aborigine Constituency. Taiwan’s indigenous population in the Lowland and Highland Aborigine districts have historically been represented by legislators from the KMT, KMT-allied splinter party, or independents.  Lastly, the third surprise was DPP candidate Lai Pin Yu, in the New Taipei City 12 District. Lai was the youngest legislative candidate in the election at 27 years old, and narrowly defeated the KMT opponent. 

Social Media Perception of the Election

On social media, DPP supporters celebrated their candidates’ victory and expressed pride in having Tsai Ing-wen as their president for another term. Several Groups and Pages shared an emotional thank you speech by defeated Legislative Yuan DPP candidate Enoch Wu to his supporters. One user shared a picture of Enoch Wu doing a deep bow to thank his supporters, a very polite gesture, commenting, “this is a real man.” Other users posted content ridiculing statements Han Kuo-yu supporters had made pre-election, such as one supporter’s declaration he would commit suicide if Han Kuo-yu did not win the election, or another supporter’s statement that he would invite all Taiwanese people to fried chicken if Tsai Ing-wen won over 8 Million votes, saying “All of Taiwan can eat fried chicken!”

Groups of Han Kuo-yu supporters reacted to their candidate’s defeat in various ways. First, several groups renamed themselves, such as 2024韓國瑜總統後援會（全球總會） (previously 2020韓國瑜總統後援會（全球總會）) changing their 2020 support for Han Kuo-yu for president to support for Han Kuo-yu in the 2024 presidential election. Other groups added the phrase “mayor” or “city government” to their name, or changed the term “president” (總統) to “mayor” (市長).

Examples for name changes; Left: Group replaces “president” (總統) with “mayor” (市長), Right: Group changes 2020 to 2024.

In groups and across pro-Han Kuo-yu/pro-KMT pages, fans expressed feeling sad or depressed about the election result, saying that their “heart hurts” and writing in one post, "Mayor Han you don't need to apologize to the Taiwanese people, the Taiwanese people need to apologize to you!" Others claimed to be bullied by DPP supporters, sharing a video of a Kaohsiung middle school student using a Han Kuo-yu doll to wipe a chalkboard, saying that such poor behavior represented the overall “demeanor of the green people” (green being the color of the DPP). In one popular post, a user protested “Green Terror” and stated “Taiwan was not saved.” Lastly, a post that received 7,000 likes and 1.4 thousand comments complained about allegations against China-friendly news outlets such as China Times. The post asked Han supporters if, should the pro-China outlets be shut down, they would stand up and defend their rights to watch such content.

Image

Additional battlegrounds - YouTube

While Taiwan’s election is over and there was no interference attributed to Beijing, a statement from Chinese officials on Sunday reflected continued tension: “No matter what changes there are to the internal situation in Taiwan, the basic fact that there is only one China in the world and Taiwan is part of China will not change.” To that effect, Taiwan is concerned about ongoing attempts by an external force to spread disinformation and control the narrative online. One battleground that appears to be of increasing concern as a potential site for creating and sharing false information is YouTube. A recent article by TechCrunch highlights that over the past two years, YouTube has risen as a “potent” way to spread disinformation in Taiwan. Several researchers in Taiwan have raised this concern, including before the election.

In a letter to the editor published November 2019 in Liberty Times Net and quoted in English in its sister newspaper, Taipei Times, Wang Tai-li (王泰俐), a professor at National Taiwan University’s Graduate Institute of Journalism, wrote that because Facebook and Twitter have shut down fake accounts originating in the PRC (focused on Hong Kong), platforms such as YouTube have become more of a threat, where the channels can disseminate “false information and plausible images” on a massive scale. While this is not a foregone conclusion - YouTube has also taken down suspicious activity related to Hong Kong - Wang highlights several specific examples to illustrate this concern. One suspicious YouTube channel highlighted by Wang and reviewed by the Taiwanese Bureau of Investigation is 希達說台灣—玉山腳下, translating in English as “Xida speaks on Taiwan - At the foot of Yushan,” where Yushan is the tallest mountain in Taiwan. This channel mainly focused on denouncing Tsai Ing-wen’s government, with one popular video published in September 2019 questioning whether Tsai actually received her PhD from the London School of Economics (LSE). LSE confirmed her degree in a statement in October 2019 in an attempt to dispel the rumor. In October 2019, the Bureau of Investigation revealed that the host of the channel was a Chinese state reporter. Since the Bureau’s announcement, the channel has stopped uploading videos. However, the channel is still up, which means, as Wang argues, that anyone can view the videos without knowing the information is false. Wang’s research discovered 10 similar YouTube channels to “Xida speaks on Taiwan - At the foot of Yushan” that were created from August to October, some of which are still up on YouTube. According to the Taipei Times, Wang’s research finds that some of the channels with more than 10,000 subscribers are content farms run by Chinese nationals and have simplified Chinese subtitles, which are used in the PRC but not in Taiwan.

Our team at SIO noted the presence of YouTube channels with anomalous behavior that became popular sharing content on other platforms during the recent election. Several of these YouTube accounts were created years ago, but only appear to have begun to post videos recently. Several have unusual growth and viewing patterns. One of these YouTube channels, 高雄林小姐 (“Kaohsiung Miss Lin”) was created in December 2011, but only began to post videos in October 2018, one month before the local elections in which Han Kuo-yu was elected as mayor of Kaohsiung. In the first posted video, the speaker says, “I am not a ‘Wumao’” (Wumao is a term commonly used to describe Beijing’s 50 cent army – bots and trolls paid by the PRC government) and then goes on to claim that Tsai Ing-wen, in fact, uses an “online army.”

Image

Most recently, Google has taken action to address false information on their platforms specifically in Taiwan. Ahead of the 2020 elections, Google published an article on its Taiwan-Official Blog about a recent annual Media Summit in Taiwan where various fact-checking organizations and newsrooms came together to discuss best practices in fact-checking, increasing media literacy, and responses to false information. The summit focused on bringing a host of stakeholders and researchers to one place to share their latest methodologies for fact checking and flagging false information. The objective of this program, the Google News Initiative writes, is to support the media workers and professionals who are bringing  “the latest information back to Taiwan’s applications. We also intend to continuously implement more cooperation and exchanges in Taiwan to combat false information.” (translated).

Takeaways from our research

In our Taiwan election scene-setter blog post, we hypothesized about two prongs of influence operations that might be leveraged in Taiwan’s 2020 election. The first was media manipulation: concern that Beijing might facilitate strategic persuasion and inject disinformation narratives into Taiwanese local media. The second type of interference we hypothesized might be present was subversive social influence using fake Pages and accounts, similar to the significant state-backed information operation targeting Hong Kong protestors on Twitter. 

Over the past five months, we did not find any cases of disinformation on social media that we believed to be attributable to the PRC. Facebook, while it did a takedown during the campaign, did not attribute those Pages or Groups to outside influence. We did note some suspicious activity, however, most of it appeared to be domestic hyper-partisan fan groups. 

On the media front, we observed the spread of Beijing’s propaganda via their persistent control of channels of influence within the media ecosystem. This observation has been echoed by many, including Cedric Alviani, director of the Asia bureau of Reporters Without Borders, who, in a Foreign Policy article published after the election, accused China of “exploiting a weakness” in Taiwan’s media environment. The article described the media environment in Taiwan as “long a hotbed of poor journalistic standards as outlets rush to be first to publish.” 

From our monitoring of Facebook Pages and Groups, we noted how key stories, such as that of the disillusioned PRC spy-turned-defector, have been amplified and weaponized on both sides to discredit the other party. Validating the truth of the story is second to the effort of simply making the story serve the political goals of one’s own side This partisanship has resulted in an environment in which people believe or disbelieve a story based on how well it conforms to preconceived biases, and how well it fits the needs of their political tribe.

Hyper-partisanship is detrimental to fighting disinformation. In Foreign Policy, Nick Aspinwall describes a polarizing effect in Taiwanese politics: DPP supporters are fixated on the fear of Chinese influence, while the KMT denounces this fear as overblown. The partisan divide colors the incorporation of all emerging stories - even false ones - in national narratives. Fierce entrenchment minimizes the desire to truly verify allegations of Chinese meddling before these stories are turned into talking points by Taiwanese politicians. 

While there does not appear to have been a PRC influence campaign on Facebook, this election was important to observe not just for Taiwan but also to understand what other countries facing possible external influence operations via Beijing might experience. A Freedom House special report from January 2020 found that China has increased its influence abroad. In the same interview with Brian Hioe of New Bloom, Puma Shen (沈伯洋) spoke about the importance of collaboration in detecting Chinese state-sponsored influence, believing that Taiwan’s experience contending with disinformation might be useful for other countries. In sharing his and other researchers’ analysis, Shen says, “our hope is to do this so that we can help other countries resist disinformation attacks from China and the CCP.” This quote from Shen highlights the need for researchers studying disinformation to collaborate with one another - and to raise the bar of evidence and degree of scrutiny - in order to better spot the influence operations that are evolving everyday.

There is only one day left before Taiwan heads to the polls, and researchers, election integrity teams at tech platforms, and press are following the dynamics closely. On January 1st, Taiwan entered into its ten day polling black-out period, a time during which there is a strict ban on agencies and individuals sharing, or citing, any public survey related to a candidate or the election overall. The day before the blackout period took effect, popular polls continued to show incumbent President Tsai Ing-Wen leading with a high margin.
 

A snapshot of The New Lens presidential poll shows Tsai Ing-wen with a clear majority over her opponents. 

Since the blackout period began, two stories have appeared in news coverage and social chatter, sparking rumors, misinformation, and sustaining already-present concerns about disinformation and election meddling. The first story is an update to the saga of the disillusioned People’s Republic of China (PRC) spy (covered by SIO here), Wang Liqiang, who defected to Australia in November of 2019 and claimed that he had run disinformation campaigns targeting Taiwan, specifically to support Kuomintang (KMT) candidate Han Kuo-yu; emerging information seems to implicate Taiwanese politicians and figures in a scandalous silencing attempt. The second story originates in the city of Kaohsiung, where misinformation about a Democratic Progressive Party (DPP) parade is competing on social media against the fact-check that debunks the rumor. 

An Update to “The Disillusioned  Spy” Case

On January 8, 2020, The Age, the Australian paper that broke the story of a self-proclaimed Chinese spy-turned-defector, reported that Australian intelligence agencies had learned of threats to the lives of Wang Liqiang and his family if he did not recant his story. The threats demanded that Wang record a video stating that he had been bribed by the DPP of Taiwan to make false claims about spying -- or he would be sent back to the PRC and killed. The Age article names two individuals behind the threats: a PRC-based businessman known as Mr. Sun, and the current deputy secretary of the KMT, Alex Tsai. These accusations have sparked a response in both media and social channels. Much as the original story was leveraged by the DPP and KMT to reinforce support for their own party and denounce the machinations of the other, the media and social media responses to the latest developments are split along party lines, either supporting Wang’s credibility or discarding the allegations entirely as a “spy farce.” 

The story has had a significant impact. After months of deliberation and opposition from some KMT legislators, the Taiwanese legislature passed the Anti-Infiltration act,(反滲透法) on December 31st, 2019, which aims to curb PRC influence in Taiwan’s politics. While this was not the first time legislation of this type was considered, according to a Reuters report DPP lawmakers pointed to Wang Liqiang’s allegations to explain increased urgency for passage of the bill.

Coverage in the Media

Hours after the new details surrounding Wang’s story re-surfaced, more pro-KMT news outlets such as United Daily News asserted Wang’s claims were still not credible. In one article, Wang’s story was referred to as “too outrageous”, with insinuations that he always appeared in the media at sensitive times. Another UDN article responded to the allegations by decrying “black hands” (a Chinese language term referring to evil-doers) were behind the allegations and condemning the DPP for creating the scandal to slander the KMT. 

More ‘green-leaning’ (pro-DPP) news outlets like Liberty Times News, however, reposted the issues raised in Australian media, stating that KMT politician Alex Tsai had demanded that Wang claim he was bribed and his actions were instigated by DPP. In the PRC, state media outlets such as Global Times have continued to discredit Wang’s claims by denying his identity as a spy and portraying him as a swindler, writing that the West has been “hyping” up allegations of PRC efforts to meddle in the Taiwanese elections and citing Wang’s story as an example. 

Global Times coverage of the Taiwan election, meanwhile, has remained largely focused on general discussions about Taiwan’s preparation for its upcoming election, vaguely alluding to “accusations of mainland meddling and influence of Hong Kong’s social unrest,” and stating that PRC mainland officials and analysts believe that unification “of the motherland” is “an inevitable trend regardless of who wins.” 

Engagement on Social Media 

When the disillusioned spy story initially broke in November, the SIO team began to track a number of channels and videos on YouTube discrediting Wang Liqiang as a liar; these videos were widely shared across pro-KMT Pages on Facebook. 

In one video shared across various Han Kuo-yu supporter groups (79,131 views on YouTube and 21,400 interactions on Facebook), the speaker discredited both Australian media (“They have no responsibility, Mainland sources include the Public Security Bureau and court. Who has credibility?”) and Wang (“Is he a scammer?”) whilst stating that Han’s popularity is purely organic (“Are millions of people brainwashed?”). Several top comments on this video on both YouTube and Facebook are written in simplified characters from accounts that list the PRC as their location. Other observers of the Taiwan election have noted highly polarizing viral YouTube activity as well; a recent piece by the New York Times discussed similar YouTube accounts that shared content aimed to “inflame and divide.” 

Most recently, the New Party (politically close to the KMT) live-streamed a KMT press conference discussing Wang Liqiang on YouTube and Facebook. During the livestream, live chat was deactivated, but before it was, several commentators decried the “1450” (alleged DPP cyber army, which Han Kuo-yu supporter call “1450”) for “being at work”. Many comments were written in simplified characters, which are used in the PRC Mainland but not in Taiwan.

Comments in the YouTube Live Chat (Top chat replay) allege the 1450 is attacking the comment section, e.g. ​1450上班了 (“1450 is at work”); several comments are written using simplified characters. 

KMT- and DPP-support groups also disseminated content relating to the disillusioned spy story. Within the KMT groups, various users shared articles (almost exclusively from China Times) commenting that the DPP was accusing other parties of being misleading despite being the most misleading party itself. Within pro Han Kuo-yu fan pages, posts about Wang’s new claims were accompanied by articles from China Times. One article that appeared in several pro-KMT groups claimed that Tsai Ing-wen government was actually behind Wang’s story, and that a DPP legislator, Chiou I-jen (邱義仁), went to Australia to visit Wang and give him money. After the circulation of these articles, however, KMT politician Alex Tsai goes back on his claim that Chiou I-jen pressured Wang to falsify his story.    

The Age article, though shared by several pro-DPP pages, did not appear to be shared in any of the public pro Han Kuo-yu Pages or Groups. DPP supporters ridiculed the KMT for alleging that the DPP was behind the disillusioned spy story. 

Image

Image

Parades through Kaohsiung spark misinformation: a look at the dynamics of fact-checking 

On December 21st, two major parades took place in Kaohsiung. The first, referred to in the media as 罷韓、(“Stop Han”), or the “1221 Wecare Taiwan Parade” took place in the southern part of the city and was a mix of Tsai Ing-Wen supporters as well as civil society organizations such as Wecare Kaohsiung, Step on the Red Media, and Support Hong Kong (translated). The Stop Han parade was framed as a march to ‘reclaim’ the city of Kaohsiung from current mayor and KMT presidential candidate, Han Kuo-yu. The second parade, 挺韓 (“Support Han”), consisting of Han Kuo-yu supporters, took place in the northern part of Kaohsiung. The parades generated heated conversations on social media, as rumors about attendance - and then messages to rectify those rumors - spread across partisan fan communities, including Facebook Pages and Groups. This case provides us with an interesting opportunity to see not only how misinformation spreads on Facebook, but how the fact-check traversed the community as well.

One of the false rumors circulating about the parades claimed that the Democratic Progressive Party had used a fake photo of their parade to make turnout look greater than it was. Eight Han Kuo-yu fan Pages each posted the same message discrediting the photo, disparaging it as a fabrication, within two minutes of each other (some mere seconds apart). The posts were individually created and used verbatim text; they were not Shares of the same post originating on one page. This close posting timeline may suggest that Pages or admins coordinated to spread the allegation.  

Image

The Taiwan Fact Checking Center, a Taiwanese nonprofit aimed at fighting disinformation, verified that the photograph was in fact real. The Center both geolocated the image to the street in Kaohsiung, and interviewed the photographer, who provided the Center the original photo on his camera. The article debunking the rumor was shared on Facebook by the Fact Checking Center and by other outlets such as Wecare高雄. 

Engagement and Reach of Both the Fake and Fact-Checked Stories on Facebook 

Image

The table above conservatively documents how both the rumor and the fact-check about the parade image played out on Facebook. We examined the reach of the fake parade picture by searching for key terms that describe the image, and snippets of text from the post, across public Facebook Groups and Pages. Then we located the factcheck - first posted to the Taiwan FactChecker Center’s website - and searched for shares and reposts of the correction in public Groups and Pages. 

The fact check received three times the number of shares as the posts alleging that the image was fake, and received more interactions (comments, Likes) in aggregate. Interestingly, the Taiwan Fact Checking Center assessment was shared into two pro-Han Kuo-yu Groups. Most of the comments on the pro-Han Kuo-yu shares of the factcheck post, however, continued to contest the validity of the image (and credibility of the Center). 

Image

In the post detailing the fact-check on Taiwan Fact-Checking Center’s own Facebook Page, the Facebook comments are primarily positive; many appear to be DPP supporters who believe that the fact-check reinforces their message that the Han Kuo-yu camp is dishonest and can’t be trusted. Some of the comments mock Han Kuo-yu supporters or Han Kuo-yu himself. One commenter mocked the need for the fact-check in the first place, saying “Korean fans have no common sense and no knowledge, only know that they often watch Zhongtian TV. No wonder they have been making jokes.” 

Conclusion: What to watch for on Saturday

Although the crowd-size and photo-veracity debates seem trivial compared to more pressing issues in the election, the stories are emblematic of the misinformation and rumors actively circulating ahead of Saturday’s election. 

According to the Central Election Commision, the blackout period ends at 4pm on Saturday January 11, 2020. Experts following the race closely have predicted that Tsai Ing-Wen will be re-elected, which portends a reassessment of the PRC’s strategy towards Taiwan. Although the presidential election looks to be decided, there are still toss-up races in the Legislative Yuan that observers are closely monitoring, such as those in New Taipei and Taoyuan.

As all eyes watch for the results to come in on Saturday, the government of Taiwan and organizations such as the Taiwan Fact Checking Center are on the lookout for disinformation. Audrey Tang, Digital Minister of Taiwan, who has been leading efforts to combat false information, has worked to keep Taiwan informed about the challenges of influence operations as people head to the polls tomorrow: “We understand that the people who are sowing discord are also building a community, that they are also learning from each other’s playbooks,” Tang said in a recent New York Times article “There are new innovations happening literally every day.” 

We thank Dr. Kharis Templeman for providing helpful feedback and background information on this post.

This post has been updated on 4/29/20 with a correction that the same message was shared across eight, not seven or nine Facebook Pages.

Last Friday, December 13, 2019, Facebook announced it had removed 118 fan pages, 99 groups, and 51 accounts supporting Taiwan’s KMT presidential candidate, Han Kuo-yu. Our team at SIO had been observing several of the Groups removed, including one that was prominently featured in media coverage of the takedown: 2020韓國瑜總統後援會（總會）[“2020 Han Kuo-yu presidential support group (General group)”]. In this weekly update, we discuss 2020韓國瑜總統後援會（總會), look at how the news of the takedown has affected the Taiwanese political social media conversation, and discuss a specific misinformation campaign related to voting.

2020韓國瑜總統後援會（總會）[“2020 Han Kuo-yu presidential support group (General group)”] 

The 2020韓國瑜總統後援會（總會）group was created in September 2017, and went by the name “Kaohsiung Fan Group” before Han Kuo-yu ran for mayor. At the time of its removal, the group had over 150,000 members and 109 admins and moderators - unusually high compared to the average admin and moderator counts for Taiwanese political groups of either affiliation (pro-Han Kuo-yu groups averaged 27, and pro-Tsai Ing-wen Groups, 10). 

Several of the moderators for this Group were suspicious: the profiles had zero or one friend, profile pictures that were not of real people (ie, pictures of a Chinese landmark or an actress), and minimal human engagement on their posts, photos, etc. These accounts appear to have been removed as part of the recent takedowns.

An example of one of the suspicious admin profiles that has since been taken down. 

Content-wise, 2020韓國瑜總統後援會（總會）focused on calling on members to vote for Han Kuo-yu, raising awareness about the candidate’s activities, and quoting sources and press that claim that Han Kuo-yu will win. One post from December 5th stated that Xie Longjie (謝龍介), a KMT politician, estimated that Han Kuo-yu would win the election by a margin of 900,000 votes.

Reaction to the takedown in across social media groups 

Reaction to the takedown has sparked a collective response from members’ whose pages got taken down, and initiated an attempt to regenerate  their groups. A new group, 2020韓國瑜總統後援會（全球總會）, which added, “全球” to its name, which translates to “international support group” promoted within various other pro-KMT groups as a replacement for the removed 2020韓國瑜總統後援會（總會） already has 17,054 members as of today and shares several of the administrators of the defunct 2020韓國瑜總統後援會（總會). 

Image

​2020韓國瑜總統後援會（全球總會) membership as of December 19, 2019.

Another Group, formerly 韓國瑜總統 唯一支持韓國瑜 (“President Han Kuo-yu Only support Han Kuo-yu”), renamed itself 2020韓國瑜總統後援會（總會), most likely to opportunistically benefit from the media and social media attention on the removed group of the same name. However, the rebranded group is being decried as a “copycat”, and activists are encouraging their audiences to join the new replacement instead.

Image

Image

The removal of the pages, groups, and accounts has also prompted conspiratorial theories about why they were taken down in the first place. Many posts in the remaining network of KMT Groups decried Facebook’s efforts as being in coordination with the DPP (the party of presidential incumbent Tsai Ing-wen), for example sharing this video in which plays text of a picture of Mark Zuckerberg and talks about the takedown. Online activists and group members described the efforts as censorship and even shamed it as a new “Green Terror” (“綠色恐怖”). This is a reference to the White Terror - the brutal oppression of political dissident Taiwanese people endured when the country was a dictatorship - rebooted to allude to the DPP, whose party color is green. 

Image

Image

Some DPP-linked fan pages also discussed the removal, although the topic was secondary to a continued focus on the Hong Kong protests. DPP supporters leveraged the takedown to decry the hypocrisy of ongoing KMT allegations that  the DPP uses online bots and trolls to promote its position.

Overall, the reaction to the takedown is similar to the partisan reaction we have observed in  response to other recent news, including last week’s blog post in which we discussed the KMT accusation that the DPP had a green “net army” used to attack the diplomat in charge of the Kansai International Airport evacuation. In both cases, KMT-aligned supporters spoke about how their group is the victim of untrue stories--in the green “net army” that the DPP is ironic in accusing the KMT of influencing others online during the 2018 election, and in this most recent case by saying that Facebook is aligned with the DPP.  

Other anomalous behavior: a manipulative meme about ballot numbers

Image

A final observation related to activity surrounding the Taiwan election in Facebook communities relates to memes about voting. On December 9th, the presidential candidates drew numbers that correspond to their placement on the ballot. KMT candidate Han Kuo-yu and his vice president ticket got  number 2, while Taiwan’s current president Tsai Ing-Wen’s ticket chose number 3. These numbers were then incorporated into catchy slogans by each campaign,designed to remind voters to vote for their candidate. Later that day, the Taiwanese publication Taro News reported that a KMT-fan Page had posted a picture with a cartoon of Tsai Ing-wen and her running mate with the caption: “Taiwan wants to win, vote No. 2”. The picture shows a cartoon Tsai and vice president holding up a number two. The post has since been removed, but DPP supporters have made a YouTube video and spread screenshots to create a record that it happened. People angry that KMT supporters would play dirty and mislead fellow voters have left comments and screenshots of the misinformation meme on nearly every post on the Page since the incident. In the screenshot above, the red arrows point to the misinformation, and the comment reads, “As a Taiwanese, I feel that this behavior is really shameful. People must have ambitions. A party with many ambitions has no credibility.”

Looking forward to 2020

We have observed a range of behaviors within fan groups for both the DPP and KMT party, including Page/Group name changes, identical posts blasted across multiple fan pages within a few minutes by the same account, and profiles dedicated exclusively to political activism (with few indications that they are used for more general Facebook activity), but they all appear to be domestic and show no indication of foreign interference. Memes containing misinformation appear, but gauging the extent to which they are intended to deceive (and their original provenance) is at times difficult. This reflects the challenge of placing activities on the spectrum from legitimate activism to manipulative behavior. We will continue to watch the engagement and behavior of these Groups and Pages as the election nears. 

This post is an update to our Presidential Election 2020 Scene Setter published August 26, 2019. 

Update (12/13/19 3:00pm): a few hours after we published this blog post, Facebook took down 118 fan pages, 99 groups, and 51 accounts for inauthentic  activity and rules violations. The story is developing , but as reported in the press the issue appears to have been inauthentic behavior designed to artificially inflate the popularity of content in certain communities, rather than foreign influence. The Page names were not officially released, but it appears that one of the largest pro-Han Kuo-Yu groups, 2020韓國瑜總統後援會（總會）, is now down. This Group was one of dozens of extremely active political fan communities that we had examined, noting coordinated posting, regular "overperformance" of content in CrowdTangle, and Page manager overlap with other pro-KMT communities beginning in early November. This takedown illustrates the complex spectrum from coordinated activism to coordinated inauthentic activity, whether domestic or foreign. We will post more about these Pages, including behavioral and content observations, early next week.

The Taiwanese presidential and legislative elections are now one month away, and the SIO team has been following emerging narratives. The candidates in serious contention are Democratic Progressive Party (DPP) President Tsai Ing-Wen and Kuomintang (KMT) candidate Han Kuo-yu. Currently, Tsai holds a 20-35 point lead over Han depending on the poll. According to experts who are following the election closely, this past month was a turning point in the election where Tsai Ing-Wen looks now to be in good shape to win reelection and the odds for the DPP holding onto the majority in the LY appear better as the election nears. Political and economic factors such as the connection the Taiwanese feel towards the Hong Kong protestors, Tsai’s firmer stance towards the PRC, and a healthy economy have all played in Tsai’s favor.  Additionally, despite previous concerns about online influence operations by Beijing to boost Han Kuo-yu during his 2018 mayoral campaign, and calls from the government to be wary of PRC interference, we have not yet observed coordinated inauthentic behavior supportive of either candidate in the 2020 presidential campaign. 

Nevertheless, in this period two notable polarizing stories have dominated election coverage recently, which the competing candidates and their supporters are framing in a way that discredits their opponents. First, confessions of a PRC-run disinformation campaign in the November 2018 election, made by an alleged PRC operative who recently defected to Australia, has excited pro-DPP fans as proof of the PRC’s meddling on behalf of the KMT party. On the other side, new information has surfaced from a year-old disinformation campaign surrounding the crisis of Taiwanese tourists trapped at Kansai International Airport in Osaka, Japan due to a typhoon and is being leveraged by the KMT as proof that the DPP has a green “net army” (green being the color of the DPP). Combined, these stories have introduced public confusion about what can and cannot be believed leading into the elections.

The “Disillusioned Spy” - In the Media

The story of an alleged Chinese defector broke on Saturday, November 23rd in Australian media outlets. In what was billed as a world exclusive interview with 60-Minutes Australia, Wang “William” Liqiang, reported disinformation efforts he managed on behalf of the Chinese Communist Party (CCP). 

In interviews, Wang recounts working for a senior intelligence operative at China Innovation Investment Limited –  an alleged cover organization for PRC intelligence agencies and Communist Party officials, he claims. While at this company, Wang worked to thwart democracy movements in Hong Kong by collecting intelligence on pro-independence activists. He recruited mainland students, “influenced them with patriotism,” and sent the students to collect personal data on pro-independence student activists. This data was then made public, putting the targeted students’ family members at risk.

In addition to these claims, Wang also professed to have meddled in the August 2018 elections in Taiwan. The Financial Times reported on Wang’s allegations that CCP-backed operatives created 200,000 fake social media accounts and 20 internet companies to attack the DPP through operations online. Wang also claimed that RMB 1.5 billion was given to Taiwanese media companies to promote Han Kuo-yu’s campaign for Kaohsiung mayor. 

Within Taiwan, Wang’s accounts of cyber espionage were the subject of media coverage as well as extensive social media activity. Taiwan’s four primary daily newspapers - The China Times, United Daily News, Liberty Times News, and Apple Daily - covered the story, though in somewhat different ways. 

China Times and United Daily News - outlets considered friendly towards the PRC and more blue (KMT) leaning respectively -  commonly described the story as a “spy farce.” The publications worked to discredit Wang and his claims, reporting that his accounts were “highly suspect” and suggested he was a “mediocre spy.” PRC state media such as China Daily and Global Times shared similar rhetoric, with the China Daily writing an editorial titled “Australia's spy stories defy belief.”

By contrast, Apple Daily uses more neutral language to describe the developing controversy. They reported, for example, that while some details are difficult to confirm, Western diplomats are taking Wang’s claims seriously. In a more recent article on December 7th, the Liberty Times News shared a similar story, citing people in the US intelligence community, such as PRC expert Nicholas Eftimiads, who believe Wang’s claims.

“Disillusioned Spy” Coverage on Social Media

Partisan Facebook Page and Group conversations are also sharply divided in their perceptions of the “disillusioned spy.” Our team tracked the themes and stories that have received heightened attention as each party shared the side of the narrative most favorable to them. 

A search for Wang’s name (王立強) across a subset of public pro-DPP Fan Pages on Facebook shows that popular posts have focused on: 

• Denouncing a video released by Chinese state media claiming to show Wang appearing before a judge for fraud charges in 2016. An example of this story can be found here.
• Rebuking Han Kuo-yu’s claims questioning the credentials of the Australian journalist who interviewed Wang. An article that received high engagement on Facebook titled, “No YOU’RE the small toothpick,” defends the credibility of the journalist after Han allegedly described him as a “small toothpick.” 
• Circulating Wang’s claims of donating money to Han Kuo-yu’s campaign in 2018. Han has been quoted numerous times saying that if he had received a dollar from the CCP, then he would have immediately withdrawn from the election. One pro-DPP Page writes in a Facebook post, “According to his previous logic, it should be quite a lot of money.”

Across a subset of public pro-Han Kuo-yu Pages, narratives have focused on: 

• Defending Han Kuo-yu against Wang’s corruption allegations by saying that Tsai is playing dirty to win. One Facebook post for example links to a China Times article defending Han Kuo-yu against Wang’s corruption allegations. The post also showed a photo of Tsai Ing-Wen with a caption stating that Tsai lacks evidence for the allegations and is “shameless” in her attempt to win at all costs.
• Claiming that the DPP is using Wang’s story of CCP intervention as a “smear to change election results.” The quote is from an article reported by Storm Media that received about 14,000 likes and 1,200 shares on Facebook.
• Dismissing Wang as someone not high up enough in the ranks to listen to. Han Kuo-yu himself weighed in on the news story, posting a link on his Facebook page to an article that links to coverage from Sky News Australia and The Daily Telegraph where Australian “high-ranking sentimental officials” determined that Wang was too young to be a high-level spy. 

On Twitter, the Epoch Times played a prominent role in circulating coverage of Wang, taking a stance in their articles defending his claims from attempts of PRC-state media to discredit him. The US-based publication was founded “in response to communist repression and censorship in China.” In the United States, the American-based publication prides itself on presenting an alternative to mainstream media coverage. Through additional regional offices in China and Japan, Epoch Times spread Wang’s story through the hashtag, “#王立強” (“#WangLiqiang”). 

Image

Kansai Airport Flooding Story Resurfaces

Less than two weeks after the story of Wang’s defection, another disinformation-related story broke - an unexpected update to a debunked fake news story which circulated in early September, 2018. Last year, the Kansai International Airport in Osaka, Japan flooded due to a typhoon and an evacuation was needed for travelers flying through the airport. On September 6th, PRC state media reported that Taiwanese travelers had to identify as Chinese in order to be allowed on the bus that was used during the rescue efforts, among them “32 Taiwan compatriots” a China Times article wrote. On the same day, a thread about the evacuation was created on a popular Taiwanese online forum, Professional Technology Temple (PTT), and users posted links to the PRC media articles. Newspapers in Taiwan and Hong Kong picked up the story, reusing quotes in a Global Times article from alleged witnesses. These news articles were later proved to be false, as witnesses came forward to state that this had not actually occurred. An investigation found that the IP address of the creator of the Kansai airport thread and the accounts posting fake news stories were traced to Beijing, and so the operation was assessed to be a PRC influence operation.  

At the same time the fake news stories were being shared, another PTT account with the username “idcc” posted scornful messages criticizing the Taiwanese consulate for failing to protect its own citizens. On December 2nd the story resurfaced: posts attacking the consulate were attributed to Tsai Fu-min, who Taiwanese media described as an “internet troll” paid by DPP-linked figure Slow Yang (楊蕙如). Though Yang denies the charges, she was indicted by the police. The DPP is trying to distance itself from her, stating that she was not acting on the orders of anyone in the party. 

In another layer to the intense feelings and misinformation around the evacuation from Kansai airport, a week after the flooding, the director, Su Chii-cherng (蘇啓誠), in charge of the Taiwan representative office in Osaka and the target of the PTT criticism, committed suicide a week after the flood. Taiwanese media outlets reported that in the director’s suicide note, he cited the humiliation of being criticized online. Our team cannot independently verify these claims and have found articles citing other reasons for taking his life.  

The new development linking Yang to the story raises questions about the previous attribution to Beijing. Moreover, the story has stirred a lot of interest among the pro-KMT Facebook Pages, who are using the story to discredit the DPP. One common theme is comments and posts demanding to know where Yang received her money to fund the “net army.” Some posts state that it’s ironic that DPP is accusing the PRC of being involved in the Taiwanese election when they have a “green net army.”

The fervency of this story is not just among voters but also leveraged by Han Kuo-yu, who posted about the arrest of Yang and the tragic death of the Osaka consultant director. In his post, Han Kuo-yu criticizes the DPP’s place in the story, calling the party “cold blooded.” In their own response, on December 2nd, the DPP issued a statement on Facebook that Yang’s case has nothing to do with the DPP and condemned the KMT for “deliberately linking this matter to the DPP.” 

The story still continues: in Taipei on Deceber 6th, a group of KMT legislators and city counselors attempted to confront the Foreign Minister, Joseph Wu ((吳釗燮), over the suicide of the Osaka consulate director. The Taipei Times and other news outlets reported that, blocked by police, the legislators used what the DPP accused as “excessive violence” against the officers. The paper then reported that two of the KMT legislators ended up going to the emergency room, where Han Kuo-yu visited them. Han’s visit initiated another controversy, the headline of the article quoting the Taipei mayor saying, “hospitals are no place for politics.”

Conclusion - What to watch over the next 31 days

While we have seen Tsai Ing-Wen’s numbers rise by 2 percent and Han Kuo-yu’s numbers drop by 6 percent the week after the news of Wang’s story broke, we can’t make the claim that these stories are impacting the average swing voter. Rather, what seems to be more likely is that these narratives reinforced the entrenchment of these online groups committed to either the DPP or the KMT. 

That said, there are three narratives to watch over the next few weeks. The first is the ongoing attempt by the KMT to discredit DPP support as being the product of a cyber army. If the KMT has low expectations of winning  the presidency, this narrative could allow them to attribute a Tsai Ing-Wen victory to unfair means. 

A second story to watch just broke yesterday on December 11th: the Criminal Investigation Bureau (CIB) detained 10 individuals accused of enabling more than 10,000 “tourists” over the past three years to bypass official background checks when entering Taiwan. Included in this list are individuals known to investigators as being involved with the United Front, a popular front of political parties led by the CCP.  The person allegedly running the operation, Hung Chin-lin (洪慶淋),is a former KMT county-level official, though he has not worked for the party for more than ten years. 

The third thing to watch are the Legislative Yuan elections. The legislative elections are of interest to entities in Beijing looking to influence Taiwanese politics, and if the presidency is no longer in reach, that interest may increase. The DPP currently holds 68 of the 113 seats in Taiwan’s unicameral parliament (holding an 11 seat edge). Until recently, analysts and the Taiwanese media expected there to be more struggle for the DPP to win a majority as it had done in 2016. However, currently, with the polls showing an expansive gap favoring Tsai, some experts believe that the DPP could possibly add more to its majority. 

We thank Dr. Kharis Templeman for providing helpful feedback and background information on this post.

This is the second of a series of pieces we intend to publish on societies and elections at risk from online disinformation. This election monitoring project is somewhat unique: the elections in Libya may not actually happen. We choose to focus on Libya as regional actors are largely unconstrained in their ability to meddle in the country; it is an important test case to see how far malign influence operations can extend.

Libya’s UN-backed Government of National Accord (GNA) has proposed scheduling presidential elections for the end of 2019, though it is unlikely this will happen. Over the coming months the Stanford Internet Observatory will track disinformation in Libya, observing activities of myriad regional actors with both a stake in how Libyan politics unfold and a track record of running overseas social media influence operations.

Political context

Libya has seen continuing violence for several years, with rebel General Khalifa Haftar, head of the Libyan National Army (LNA) forces, controlling large swaths of the country and aiming to take control of Tripoli.

Libya became independent in 1952. Muammar Gaddafi took power in 1969 and ruled until 2011. His rule was characterized by oppression, patronage, and violence, and Gaddafi consolidated control with networks of informants and a climate of fear. In the 1990s international sanctions against the regime (for supporting terrorism) weakened Gaddafi’s patronage machine. In 2011, at the start of the Arab Spring, the Libyan people began to protest the Gaddafi regime. The government was brutal in its response, which caused further and more intense protests. With support from Qatar, the UAE, and France, rebel groups began seizing territory. In the fall of 2011, with NATO support and a US-backed no-fly zone, rebels killed Gaddafi and seized control of the government. A National Transitional Council, which included former government officials who had defected in early 2011, academics, human rights activists, and tribal leaders, took power. The Council transitioned power to an elected General National Congress in 2012.

The GNC lost legitimacy in the years that followed, primarily for failing to provide security to citizens. In response, new parliamentary elections were held in 2014 for the House of Representatives, but the results were contested. This institution’s lack of legitimacy contributed to the outbreak of militia violence later in 2014, and the Libyan government split between the Tobruk-based House of Representatives and executive leadership in Tripoli. Strongman Haftar and his House of Representatives-aligned LNA forces built support in the East. The GNA’s Prime Minister Fayez al-Sarraj maintained control in Tripoli. The most salient points of conflict between the two sides are over oil revenue and control of the Central Bank. The Central Bank is split in two, with the GNA controlling the Tripoli branch and oil revenues, and the LNA controlling an eastern branch. If the GNA were to cut off the LNA from banks, this could precipitate a national financial crisis.

There was an attempt at a peace deal in 2015; the Libyan Political Agreement, as it was known, created a Government of National Accord (GNA) but this deal never secured buy-in from the House of Representatives. In April 2019, after a series of successes in other parts of the country, Haftar began an offensive on Tripoli, bombing a migrant detention center in the process. This caused an outcry from the international community. He took some territory, but it was subsequently reclaimed by the GNA In June 2019.  

The international response to the conflict has been weak because of U.S. indifference, and disagreement among European countries about who to support. Though the US officially supports the GNA, President Trump spoke by phone with Haftar in April 2019 and praised his desire to fight terrorism.

US ambivalence has emboldened regional governments to interfere. The conflict has evolved into a proxy war, with the UAE, Saudi Arabia, and Egypt on the side of Haftar’s forces, challenging Qatar and Turkey’s support for the GNA; the GNA has incorporated elements of the Muslim Brotherhood. Both the GNA and LNA conduct drone strikes, using weapons supplied by Turkey and the UAE respectively, despite an arms embargo that has been in place since 2011.

As of October 2019 the war is at a stalemate, with neither the GNA nor the LNA open to negotiations, though Haftar recently signalled a willingness to open a dialogue. There are ongoing conflicts on the outskirts of Tripoli and in Misrata, Jufra, and Murzuq. Observers suggest that the soonest the elections could take place is early 2020, and only then if the UN can bring all factions to the negotiating table.

Potential disinformation threats

The ongoing conflict has made its way to online spaces, where all sides are spreading disinformation. This information battle is playing out primarily on Facebook and Twitter, and includes repurposing photos and videos to create false stories. For example, the Libyan affiliate of Al-Aan, an Arabic news network based in the UAE (the UAE is pro-Haftar) shared 2011 video footage of military movements, claiming they were recent. When the LNA captured a Portuguese mercenary supporting the GNA, an inauthentic Facebook account claiming to be a popular TV station pushed the untrue narrative that the mercenary was only conducting migrant smuggling surveillance for Europe. This false story went viral in Libya, and was amplified by a Saudi news network, and then The Daily Mail. 

There is compelling evidence of foreign-initiated pro-Haftar social media campaigns. In the weeks preceding Haftar’s effort to take Tripoli, there was a coordinated pro-Haftar Twitter campaign, complete with hashtags such as #SecuringTheCapital (in English and Arabic) that were amplified by bots, some of which participated in pro-UAE and anti-Qatar online campaigns in the past. On the day of the attack on Tripoli, another pro-Haftar campaign emerged with the Arabic hashtag “We support the Arab Libyan army” (#ندعم_الجيش_العربي_الليبي).  The top users of the hashtag were Twitter accounts in the UAE, Saudi Arabia, and Egypt. One account that used this hashtag had organized online campaigns for the Egyptian government in the past. Egyptian and UAE media coverage amplified the hashtag.  

There is speculation that Russia has acted to support the LNA in both the military and digital realms. However, given the uncertainty over which side will consolidate control, Russia has ongoing dialogue with the GNA, and is also considering supporting Saif Al-Islam Gaddafi, Muammar Gaddafi’s son and a possible presidential candidate. In 2018 Haftar traveled to Moscow and met with Yevgeny Prigozhin, who heads both Wagner Group, a private military company with ties to the Kremlin, as well as the Internet Research Agency, the entity involved in numerous global influence operations. Some reports suggest Wagner Group has provided mercenaries and arms to the LNA. In May 2019 Russian citizens Maxim Shugaley, Samer Khasan, and Ali Sueyfan were arrested in Tripoli, accused of attempting to meddle in the upcoming elections. They had been in Libya with Alexander Prokofiev (who escaped arrest), meeting with Saif Al-Islam Gaddafi. All three worked for the Foundation for the Protection of National Values, a Russian NGO headed by Aleksandr Malkevich, who is on the US Treasury sanctions list for spreading divisive and untrue stories on social media in the run up to the 2016 US elections. 

Findings from newly-released Twitter data

Further evidence of UAE and Egyptian ties to Libya comes from files Twitter recently released of removed accounts. Two files – one that Twitter says contains accounts originating in the UAE, and the other with accounts based in the UAE and Egypt that were managed by a company called DotDev – each contain Tweets that are supportive of Haftar, critical of the Muslim Brotherhood, and claim Qatar (and to a lesser extent Turkey) are supporting terrorism in Libya. The tweets were primarily in Arabic and English.

The first file, with accounts based in the UAE, contains 1,325,530 tweets. 2,503 of these (0.2%) contain the word Libya in English or Arabic, and 711 of the 4,248 accounts (17%) tweeted at least once about Libya. 96% of user accounts were created between 2018 and 2019, and the average account had 215 followers. The accounts claimed to be based primarily in the Middle East and North Africa. 

Image

The references to Libya begin in 2012, including one that praised the (translated) “secular liberal coalition” for gaining more votes than a party affiliated with the Muslim Brotherhood in Libya’s parliamentary elections. A small flurry of pro-Haftar Tweets appeared on February 14 and May 20, 2014, days in the vicinity of two Haftar coup attempts. A few of these used the hashtag #الجيش_الليبي_يحارب_الإرهاب (the Libyan Army fights terrorism). In the years that followed, the accounts mercilessly attacked Qatar, claiming Qatar supported terrorism in Libya. “Who would want to loot the capabilities of the Libyan state, overthrow institutions and stealing oil expect [sic] Qatar?” one Tweet said. Tweets also praised the UAE; the hashtag #Libya_UAE appeared 43 times in the file, accompanying statements about UAE’s contributions to security and humanitarian relief in the country. Many tweets retweeted content from @alain_4u, the handle associated with Al-ain.com, a UAE-based media outlet.

A meme an account shared that emphasized the importance of fighting terrorism in (among elsewhere) Libya, while claiming that Qatar is not committed to fighting terrorism.

Days after Haftar attempted to seize Tripoli, accounts retweeted a news headline saying the LNA was trying to liberate Tripoli, using the hashtag DFRLab researched earlier this year: “We support the Arab Libyan army” in Arabic.

Among the accounts Tweeting about Libya, we identified two likely fake online personas. Sara Bitar (@SaraBiitar) and Bilal Hamdan (@BilalHamda). They both claimed to be based in Lebanon and had about 5,500 followers. The accounts were created one day apart on September 2017, and had the same profile photo: a Lebanese woman at a protest, which appears to have been stolen from a 2012 news article. Sara claimed to be a “Political activist with a master degree in political and economics sciences.” We were unable to find any other online presence for either Sara or Bilal. The personas shared the standard narratives about Qatar and Turkey working against the interests of the Libyan people, and retweeted @qatarileaks, an account removed in the next file we discuss.

The second file – the one attributed to DotDev, contained 214,898 tweets between 2014 and 2019. 5,671 (3%) contained the word Libya, and 176 of the 271 accounts tweeted at least once about Libya.  

"The Libyan Army launches the second phase of Operation Flood of Dignity" reads this image tweeted by @binlibyaa, and retweeted by many accounts in the data in 2019.

These tweets had a near-identical tenor to the previous file, extolling the virtues of the UAE and asserting that Qatar is tearing Libya apart. One tweet in September 2017 said: “In order to avoid competing with #Libyan oil, #Qatar conspired to kill #Gaddafi and steal the #oil #QatariLeaks #boycott Qatar”. (We discuss Qatari Leaks below.) In February 2019 tweets alluded to possible Haftar movements: “Infighting Plagues #Qatar’s Arms [sic] in #Libya  #Qatar’s militias in Libya are about to be eradicated, victories of Haftar’s national army put them in a corner #Qatarileaks”. In March 2019, one tweet claimed Qatar’s militias were nearing “total collapse.” Tweets the day after Haftar’s Tripoli offensive said things like: “#Qatar and #Turkey United against Tripoli freedom The #Libyan National Army launched a final blow to terrorism in Tripoli, Seeks to purge the country from the abomination of the Muslim Brotherhood. #Qatarileaks”. A few days later a tweet stated that “Liberating #Tripoli provokes union of terrorism.”

Accounts in this file frequently retweeted @LyOffSpokesman. The account, now suspended, appears to have been a fake account for the LNA spokesman. A different account for the spokesman had denounced it as fake, showing a screenshot with @LyOffSpokesman claiming to be the spokesman's "official account." @LyOffSpokesman was likely a handle managed by DotDev. An account in the data appears with the same April 2010 account creation date (one of the earliest created accounts in the file), and a profile descirption that says it is the (translated) "news account" – perhaps a phrase added in after being called out on Twitter – "for the lieutenant Ahmad Al Mismari, the official spokesperson for the general leadership of the Libyan armed forces."

This image was tweeted by several users, including on May 23, 2019, just weeks after Haftar’s Tripoli offensive.

One of the suspended accounts claims to have been “The Official page of General Khalifa Haftar.” It was created in 2015, and had 2,896 followers. Though the handle is anonymized, engagement with the now-suspended @HaftarOfficial appeared only in May 2019, the only month that the account tweeted in the data, suggesting that may have been the suspended account. The account tweeted official-sounding statements about, for example, being well received in a meeting with the Egyptian President. Tweets from this account received on average 305 likes and 139 retweets. One tweet received 1,323 likes.

One account removed in this file was @qatarileaks, with 70,168 followers. Qatarileaks claims to be a platform to analyze information exposing Qatari support for terrorism; in practice it is simply an anti-Qatar propaganda machine. This account was among the most commonly re-tweeted accounts in both files. In the context of Libya alone, the #qatarileaks hashtag was used 352 times in the DotDev file. With tweets like “Qatar's foolish policies have contributed to the downfall of Libya, but that was the plan all along #Qatarileaks”, (from the first file) this handle linked to qatarileaks.com, a site with almost comically negative stories about Qatar. One story headlined “Death haunts passengers on Qatar Airways” contains no information about any death associated with Qatar Airways. The administrators for the affiliated Facebook Page are located in Egypt, and its Instagram account posts memes (including about Libya) that were shared frequently in the Twitter file. 

A screenshot of Qatarileaks.com/en, as seen on September 24, 2019. 

The about page of Qatarileaks.com.

@binlibyaa is likely another account Twitter removed in the DotDev takedown. While the screen name is hashed, the profile url linked to bnlibya.com. This account tweeted from March to May 2019, and replies to @binlibyaa covered the same time period. @binlibyaa frequently re-tweeted @LyOffSpokesman and @HaftarOfficial, along with links to bnlibya.com articles. That url was registered on April 6, 2019, two days after the start of Haftar’s Libya offensive. The account’s associated Facebook and Instagram accounts are down, and another Twitter account with a similar handle, @bnlibyaa, is suspended.

Alliances and interested parties

Non-aligned interfering party:

• Italy, with a populist right-wing government, is taking a hard line on migrants from Libya. Italy has economic interests in Tripoli and Libya’s south, and a rivalry with anti-populist French President Emmanuel Macron. France is fighting terrorism and smuggling in southern Libya.

Media environment

After the fall of Gaddafi, print, television, radio, and online media outlets flourished as a result of new press freedoms. Many of these outlets are based abroad, with information about their financial backers hard to discern. Many also have strong political slants. Militias frequently compel reporters to publish stories that push their narratives. The BBC has an overview of the slants of the main newspapers, TV stations, radio stations, and news agencies/sites. 

Facebook is the most popular social media platform for both civilians and militias in Libya.

Competing factions use Facebook, YouTube, and Twitter, the most popular social media platforms, to promote their agendas. Estimates for Facebook usage range from 53% to 60% to 65% of the population. In addition, 29% use YouTube and 8% use Twitter. Instagram usage estimates range from less than 1% to 14%. Overall, 58% of the population uses the open internet.

Narrative context

Organized social media campaigns push the following narratives:

1. Claim the opposing side committed war crimes (all)
2. Suggest that Haftar is a force for stability and security for Libya and will fight terrorism (pro-Haftar)
3. Highlight the role of the other side’s foreign backers (all)
4. Emphasize successes in securing or taking territory (all)
5. The UN is failing in its goals for Libya (pro-Haftar)

Hashtags that have likely been part of coordinated inauthentic social media campaigns:

• #SecuringTheCapital / #تأمين العاصمة (pro-Haftar, context here)
• #PeopleOfLibyaWantTheLibyanArmyToSecureTheCapital (pro-Haftar, context here)
• #ندعم_الجيش_العربي_الليبي (We support the Arab Libyan army) (pro-Haftar, context here)
• #WeAreReady (pro-Haftar)

Key takeaways and risks

Libya is a complex case: it’s a country at war, with two competing governments, and seemingly extensive foreign engagement in both the conflict theatre and the media environment. Domestic groups involved in the conflict intentionally share false and misleading content on social media platforms. Evidence suggests foreign disinformation efforts have supported the LNA as well. These efforts will likely increase when elections are scheduled, as domestic groups and foreign backers jockey for power in the new government. The Internet Observatory intends to follow campaigning around Libya’s presidential and parliamentary elections as they develop.

Challenges to studying disinformation in Libya

• Minimal impartial investigative reporting in Libya makes assessing the validity of narratives difficult.  
• Discerning media outlet bias is not always straightforward, further complicating the verification process.

For further reading 

Disinformation assessments

• Removing Coordinated Inauthentic Behavior in UAE, Egypt and Saudi Arabia, Facebook
• Walking on Hot Coals: How Disinformation Is Fueling the Fight in Libya, DFR Lab
• A Twitter Hashtag Campaign in Libya: How Jingoism Went Viral, DFR Lab
• Libyan Hashtag Campaign Has Broader Designs: Trolling Qatar, DFR Lab

News coverage

• Russia’s Wagner Mercenaries Have Moved Into Libya. Good Luck With That, Daily Beast
• A Facebook War: Libyans Battle on the Streets and on Screens, New York Times

Research

• Stopping the War for Tripoli, International Crisis Group 
• At a Glance: Libya’s Transformation 2011-2018: Power, Legitimacy, and the Economy, Democracy Reporting
• Libya Social Media Monitoring, Democracy Reporting

 Regional voices

• The Libyan Center for Freedom of the Press
• Khadeja Ramali, Social data analyst and Middle East and North Africa consultant, @k_ramali

We thank Katie Jonsson, Khadeja Ramali, and Lydia Sizer for helpful feedback on this document.

The House Permanent Select Committee on Intelligence held a public hearing on Thursday, March 28, 2019, as part of its investigation into Russian influence during and after the 2016 election campaign.

The hearing, "Putin’s Playbook: The Kremlin’s Use of Oligarchs, Money and Intelligence in 2016 and Beyond” included testimony by Michael McFaul, former U.S. Ambassador to Russia and Director of the Freeman Spogli Institute at Stanford University.

Download Complete Testimony (PDF 263 KB)

EXCERPT

To contain and thwart the malicious effects of “Putinism,” the United States government and the American people must first understand the nature of the threat. This testimony focuses onthe nexus of political and economic power within Russia under Putin’s leadership, and how these domestic practices can be used abroad to advance Putin’s foreign policy agenda. Moreover, it is important to underscore that crony capitalism, property rights provided by the state, bribery, and corruption constitute only a few of many different mechanisms used by Putin in his domestic authority and foreign policy abroad.

This testimony proceeds in three parts. Section I describes the evolution of Putin’s system of government at home, focusing in particular on the relationship between the state and big business. Section II illustrates how Putin seeks to export his ideas and practices abroad. Section III focuses on Putin’s specific foreign policy objective of lifting sanctions on Russian individuals and companies.

Watch the C-SPAN recording of the testimony. 

Media Contact: Ari Chasnoff, Assistant Director for Communications, 650-725-2371, chasnoff@stanford.edu

Midterm elections pose an opportunity for hackers interested in disrupting the democratic process

Voter registration systems provide an additional target for hackers intending to disrupt the US midterm elections; if voting machines themselves are too disperse or too obvious a target, removing voters from the rolls could have a similar effect. in Esquire, Jack Holmes explains that election security experts consider this one of many nightmare scenarios facing the American voting public—and thus, American democracy itself—on the eve of the 2018 midterm elections. (Allison Berke, Executive Director of the Stanford Cyber Initiative, quoted.)