22 April 2021

Exploring Chinese Military Thinking on Social Media Manipulation Against Taiwan

By: Nathan Beauchamp-Mustafaga, Jessica Drun

Introduction

Much has been written about China’s social media manipulation in Taiwan following the 2018 nine-in-one local elections, but both Taiwanese and Western analyses have skewed heavily towards the impact of this disinformation, overlooking how the Chinese Communist Party (CCP) developed its interest in social media manipulation, its planning and preparation against Taiwan specifically, and the evolution of its tactics over time.

This article seeks to address a gap in the current policy discussion and provide evidence of PRC planning for covert manipulation of Taiwanese social media. So far, too much of the academic and policy conversation in Taipei and elsewhere has focused on the outputs of PRC disinformation (purported examples of PRC disinformation and local reporting on the consequences), instead of exploring the inputs of PRC thinking, conceptual framing, and planning and technical preparation for executing social media manipulation campaigns. While this emphasis on outputs stems in part from well-documented difficulties in attribution of inputs, it is nonetheless dangerous to overlook these PRC primary sources, because a lack of understanding of the most likely perpetrator’s thinking is a disservice to broader efforts to combat disinformation.

Too Much Focus on Outputs, Not Enough Searching for the Inputs

Unfortunately, the Taiwanese government has so far been vague about Chinese planning and thinking for interference in either the 2018 or 2020 elections. Certainly, there has been no retrospective declassified report released to the public similar to the U.S. intelligence community assessment and later Senate reports on Russian interference in the 2016 U.S. presidential election.[1] President Tsai Ing-wen warned of the spread of fake news ahead of the 2018 election, but remained vague on its potential origins (Facebook, November 14, 2018). After the elections, her administration and the Democratic Progressive Party (DPP) were careful not to draw direct links between disinformation and the election outcomes, focusing instead on the party’s shortcomings during the campaign period—likely out of consideration for Taiwan’s hyper-partisan political environment (Taipei Times, November 27, 2018).

Other Taiwanese government officials, including those from its National Security Bureau, were more forthcoming, citing the PLASSF as the main organization responsible for social media disinformation against Taiwan during its 2018 election cycle and suggesting that its personnel numbered around 300,000 (Liberty Times, November 2, 2018; Taipei Times, November 5, 2018). The Legislative Yuan’s Foreign Affairs and National Defense Committee released a report in May 2019 that touched briefly on Chinese approaches to disinformation during the 2018 election but focused predominantly on government countermeasures.[2] Little public discussion exists about the PLASSF’s actual thinking on and approach to social media manipulation.

While Western researchers have made substantial contributions to existing debates on Chinese disinformation against Taiwan, they consistently overlook and underutilize official Chinese sources when discussing Chinese social media interference.[3] Taiwanese researchers similarly do not explore actual Chinese thinking on social media manipulation, and instead only track their outputs.[4] Taiwanese researchers also continue to misstate and misunderstand “cognitive domain operations” (CDO,认知域作战, renzhi yu zuozhan), the actual PLA operational concept motivating these operations.[5]

Again, the near-impossibility in definitively attributing disinformation to a concrete origin—as well as methodological and privacy considerations, among others—is a partially understandable explanation for the emphasis on outputs as opposed to inputs (Graphika, August 25, 2020). However, this approach overlooks primary and open-source assessments from Taiwan’s sole adversary–-and the entities with the mission sets, technological prowess, and resources to pursue large-scale, coordinated disinformation campaigns. In a preliminary effort to address these deficiencies, this article takes a close look at the primary Chinese military unit targeting Taiwan for psychological warfare, Base 311 (also known as Unit 61716).

­Evolving PLASSF Base 311 Interest in Taiwanese Social Media

Base 311 is at “the forefront of applied psychological operations and propaganda directed against Taiwan,” and “functions as an operational PLA political warfare command,” according to a 2013 study.[6] As part of the PLA’s massive reorganization under Xi Jinping since 2015, Base 311 was transferred from the now-disbanded General Political Department (GPD) to the newly created Strategic Support Force (SSF). A 2018 report foresaw that the SSF’s fusion of cyber and psychological warfare capabilities could “build new synergies between disparate capabilities that enable specific types of strategic information operations (IO) missions expected to be decisive in future wars.” [7] Based on these studies, this article focuses on writings produced by the Huayi Broadcasting Company (中国华艺广播公司, zhongguo huayi guangbo gongsi), which Mark Stokes has described as likely one of the regimental-grade units subordinate to Base 311 and its “commercial” persona operating a range of public organizations with semi-transparent PLA ties.[8]

Early Awareness in 2011

A September 2011 article by an online editor for Huayi Broadcasting already recognized the growing importance of social media in Taiwanese society, including for political mobilization.[9] Framed under the idea of “cross-strait news exchange” (两岸新闻交流, liangan xinwen jiaoliu), the article provided a detailed overview of Taiwan’s social media landscape at the time and identified how social media was increasingly influential for shaping Taiwanese public opinion.[10] It made four recommendations for propaganda toward Taiwan. First, embrace social media as a new medium for cross-strait propaganda, including to leverage existing PRC platforms already in the Taiwanese market (Sina Weibo) and encourage other platforms to open in Taiwan. Second, “establish and guide microblog topics to serve the development of cross-strait relations,” including “immediately transmitting positive information [propaganda]” and to “create public opinion situations of strength.” Third, leverage “opinion leaders” to “change the direction of public opinion” as necessary, including by inviting pro-China experts to open accounts. Fourth, embrace all types of new media (such as blogs, forums, etc.) to make up for shortcomings via social media.[11]

Many of these recommendations would be embraced in China’s broader social media propaganda strategy for Taiwan, though it is impossible to know how much direct impact this early article had. For example, PRC propaganda organs use PRC social media platforms (Weibo and WeChat) to target their messaging toward the intended Taiwanese audience, including through content in Southern Min (the closest mainland dialect to Taiwanese Hokkien).[12] The PRC has also specifically targeted young Taiwanese social media celebrities as conduits for Chinese propaganda (CNA, February 21). Moreover, the reports referenced above consistently find that China is attempting to shape Taiwanese public opinion through agenda setting via a variety of means.Image: A “Cross-Strait Youth Internet Celebrity Anchor Competition” held in Xiamen in September 2020 and broadcast on Strait TV. Taiwan’s National Security Agency has noted that China has expanded the organization of training activities for internet celebrities and e-commerce live broadcasters as part of its efforts to actively use new media for propaganda purposes (Image source: CNA).

2014 Taiwanese Political Events Drive Growing Attention

A November 2014 article by personnel from Huayi Broadcasting and Voice of the Strait (Base 311’s main radio channel) took this interest a step further, but still framed new media (Internet-based media) as supporting radio propaganda, likely reflecting institutional preference.[13] Nevertheless, the article argued that new media, including Facebook and Twitter, have “completely changed the traditional propaganda style and massively influenced […] public opinion guidance.” The authors suggested that Chinese propagandists “should research and understand” the new Taiwanese generation, and “use their familiar style, vocabulary, context, coordinate ‘down to earth’ content to make [propaganda] effectively land in Taiwan.” The article suggested using Facebook and Twitter to spread “content we create and content created in cooperation […] to increase the influence and penetration” of its propaganda. However, much of the article still centered on how social media could support radio propaganda. The article’s bullish embrace of social media but grounding in legacy propaganda probably reflected debates during this time within Base 311 and perhaps more broadly across PRC propaganda organs on how best to optimize work across a breadth of evolving media.

By 2015, Base 311 personnel realized that recent events in Taiwan had proved without a doubt that social media’s growing influence over Taiwanese politics was ripe for exploitation. An August 2015 article analyzing the November 2014 Taipei mayoral election hailed Ko Wen-je’s (柯文哲) successful use of social media as a key ingredient for his victory, especially amongst the critical youth voting bloc, and provided detailed coverage of his campaign’s social media usage.[14] The article argued that social media, and the Internet more broadly, had “brought a massive transformation to political communications,” and that netizens were now leading their own political mobilization by creating like-minded groups on social media. It added that social media was breaking local political parties’ and businesses’ traditional domination of Taiwanese media as well as weakening Taiwan’s traditional two-party system. However, the authors cautioned that the DPP’s successful embrace of new media further spread their “independence” agenda, and noted that the Sunflower Movement represented a slippery slope of social media transforming Taiwanese from being specifically against the unratified 2013 Cross-Strait Service Trade Agreement to being generally anti-China.

The article added that “new media is a double-edged sword for the development of cross-strait relations,” because it brought people on both sides closer even as adversarial views persisted, so that the risk of misunderstandings grew. Overall, the article reflected a bleak view of Taiwanese democracy and its prospects for China, arguing that the chaos of the Sunflower Movement “will certainly on one level weaken the influence of Taiwan’s ‘beacon of democracy’ for mainland Westernization, and offset the outcome of the human rights-centric turn of the Kuomingtang (KMT) and DPP’s China policy.”

The authors argued that “new media provides the ideal resource on popular will and information for political communications,” noting that Ko Wen-je was the first in Taiwan to apply big data analytics to tweak his campaign messaging on social media. In retrospect, conclusions from the 2015 article appear to have been turned into actions based on reports of PRC-run Facebook groups supporting Han Kuo-yu (韓國瑜), the surprise KMT winner of the Kaohsiung mayoral race in 2018 (Foreign Policy, June 26, 2019). The August 2015 article also explains one potential PRC motivation for social media manipulation and political interference against Taiwan: social media was damaging the prospects for unification.

A related September 2015 article described the Internet as the “fuse” for the Sunflower Movement and a “propaganda amplifier,” echoing the online trend of self-creating information bubbles.[15] It recommended, “for Taiwan propaganda, we should not only create our own platforms and join hands [cooperate] with Taiwanese websites, [we] should also proactively establish sites on Taiwanese social networks, share meaningful content, [produce] personalize information, and build social media followings with distinguished meaning.” This would be reflected in China’s broader long-term strategy of cooperating with local Taiwanese media, and an emphasis on distributing content through independent small-scale new media in Taiwan, because they are trusted and can have an outsize impact at “special times.”[16]

Embrace of Social Media Manipulation After Tsai’s 2016 Victory

A May 2016 article by Huayi Broadcasting personnel represented a change of tone in Base 311’s approach to Taiwan propaganda. It argued that the DPP’s victory in the January 2016 election made Taiwanese media “more green” and “more pro-independence,” creating a “harmful Taiwanese media environment” and turning Taiwanese public opinion against China.[17] The authors wrote about the rise of pro-DPP “green” media and their belief that the DPP victory could force pro-KMT “blue” media to follow DPP views on cross-strait policy, turning the Taiwanese public against China. Echoing earlier articles’ emphasis on the growing importance of social media for shaping Taiwanese public opinion, the authors recommended “strengthening the effectiveness of propaganda against Taiwan” and suggested that propaganda should “use diagrams, icons, cartoons, and audio-visual media” in order to “weaken the propaganda color” of the information. More ambitiously, the article argued that China should “strengthen awareness of online public opinion intervention” (强化网络舆论干预意识, qianghua wangluo yulun ganyu yishi). This specifically entailed “expanding the cultivation and use of online opinion leaders,” including having mainland scholars join Taiwanese online media to “actively confront Taiwanese netizens and play the role of public opinion leaders” to “guide Taiwanese online public opinion toward a direction favorable to us.”[18]

PLASSF Base 311 Operationalizes Manipulation of Taiwanese Social Media

If there was any doubt that China was focused on the artificial manipulation of Taiwanese social media, an October 2018 article written by Base 311 and National University of Defense Technology (NUDT) researchers and computer engineers, made this clear. The article addressed the equipment requirements to conduct “cognitive domain operations” (CDO), summarized in a 2019 China Brief article as the “next-generation evolution of psychological warfare [that] seeks to use information to influence an adversary’s cognitive functions, spanning from peacetime public opinion to wartime decision-making” (China Brief, September 6, 2019). The October 2018 article argued for the importance of applying CDO to social media, but highlighted several shortcomings, including “little research on the technology and equipment for cognitive domain operations on mainstream social networking platforms,” and elsewhere explicitly listed Facebook, Twitter, and LINE as platforms that needed to be further explored. It argued:


“Speed up the research for network propaganda technology targeted toward the real-time release on social platforms, voice information synthesis technology using deep learning and other technology, online netizen sentiment trend analysis using big data analytics…strengthen the research and development of new media technology, improve the psychological warfare operations capability in the whole media environment.”

The article also advocated embracing military-local cooperation for CDO on social media, including “leveraging the advantages of local traditional and new media advantages” and to “jointly use or lease […] existing platforms and channels” as well as “purchasing or absorbing mature local” capabilities while “ensuring secrecy.” While military-local cooperation could suggest Base 311 personnel were advocating for partnering with PRC domestic actors (perhaps Taiwan-focused propaganda organs based in Fujian Province), given the broad PRC penetration and manipulation of the Taiwanese media environment, it is also possible the authors were advocating leveraging willing or ignorant Taiwanese actors and social media manipulation capabilities.[19] In summary, the article—written in the lead-up to Taiwan’s 2018 elections—appeared to suggest that the PLA should create inauthentic content (disinformation) on social media across peacetime and conflict, including deep fakes (“using deep learning”) and fake content (using natural language processing) tailored for specific audiences (“using big data analytics”). More recent Base 311 technical writings suggest an interest in using artificial intelligence to control how this content is then injected into online platforms.[20]

Broader PLA Interest in Taiwanese Social Media

In addition to Base 311’s clear focus on Taiwanese social media as a vector for public opinion guidance (manipulation), other parts of the PLA also appear to have supported this effort. To give one example, a graduate student at the PLA’s Nanjing Political Institute provided a practical guide to blend in on Taiwanese social media in 2017.[21] The author observed that “in online communities where Southern Min and Mandarin is fully intermingled, if one sentence appears that clearly carries a Northern [mainland] communication style, it would be very easy to stick out and attract other netizens’ attention, creating an invisible wall in the online communities.” Tailored propaganda targeted at Taiwan on social media should use “Taiwanese flavor,” including “actively using diction that is close to the language of Taiwanese social network communities.” Coming in 2017, this article suggests that the goal of manipulating Taiwanese social media was so pervasive across the PLA by 2017 that a graduate student could support the effort.[22]

Conclusion

In retrospect, it is clear the PLA, and especially PLASSF Base 311, prepared for and may have executed a campaign to covertly manipulate Taiwanese social media and interfere in Taiwan’s 2018 elections. This article shed light on how long the PLA—as the CCP’s “barrel of the gun”—has been interested in Taiwanese social media and focused on exploiting it for political interference against Taiwan. Social media is simply the latest and greatest way for the PLA to artificially manipulate Taiwanese public opinion.

There needs to be greater emphasis—in Taiwan specifically, but also more generally—on the stated intentions and tactical considerations of entities within the PLA and the Chinese government charged with carrying out social media interference. The hope is that this article has provided an example of the range and depth of publicly available, primary source material on official Chinese thinking on social media manipulation against Taiwan and encouraged further exploration of these materials.

A shift from an overreliance on outputs to a more balanced view that incorporates inputs can promote more substantial debates and establish a firmer foundation to inform policy discussions. Such an approach would place less weight on outcomes, which inherently assume that one party has benefited and are contentious against the backdrop of elections, instead placing more emphasis on better understanding the actual threats and how to best combat them.

Nathan Beauchamp-Mustafaga is a Policy Researcher at the nonprofit, nonpartisan RAND Corporation.

No comments: