Taiwan: The Frontline of the Disinformation Wars Shaken by a recent viral falsehood allegedly traced to a Shanghai content farm,Taiwan’s online ecosystem is repairing its democratic defense mechanisms. By Nick Aspinwall
On September 5, when Typhoon Jebi throttled Japan and forced the evacuation of Osaka’s flooded Kansai International Airport, a separate storm of unsubstantiated social media rumors began brewing in Taiwan. In a post on the Professional Technology Temple (PTT), an online Taiwanese forum similar to Reddit, a user claiming to be a
Taiwanese citizen awaiting evacuation said he had been rescued by the Chinese government, which had sent 15 tour buses and invited Taiwanese nationals onboard. The post was quickly picked up by Taiwan’s Taiwan’s Chinese-language Apple Daily and Daily and Sanlih TV News. It was followed by a widely circulated September 6 report by China’s state-run Xinhua news Xinhua news agency, agency, which said China had evacuated 32 Taiwanese citizens, citing the Chinese consulate in Osaka. Another September 6 report by China’s state-run Global Times, itself picked up by several Taiwanese news outlets, claimed that Taiwanese tourists boarding the buses were told they could only board if they identified themselves as Chinese. Taken at face value, the deluge of news items collectively exhibited Taiwan’s Taiwan’s failure to assist its citizens during a time ti me of crisis, leaving them at the mercy of Chinese goodwill. The problem: The rumors weren’t true. Government officials now believe they were the work of state-sponsored Chinese actors aiming to destabilize Taiwan. Within two hours, other PTT users had refuted the original post and traced it to an IP address registered in mainland China. The post, according to PTT founder Ethan Du, was taken down that same day. day. But false stories continued to spread throughout Taiwanese social media and news outlets, even as doubts arose when a Taiwanese tourist who had been stranded at Kansai Airport said buses could not enter the flooded airport in the first place. On September 15, the nonprofit Taiwan Fact Checking Center (TFC) debunked the story after contacting the airport, which said it had turned down an offer from the Chinese government to send shuttle buses. Their report was released just one day after Su Chii-cherng, the director-general of Taiwan’s Osaka representative office, committed suicide. According to Japanese public broadcaster NHK , Su left a note expressing that he was pained by the scathing public criticism of his office, spurred by the cacophony of false reports, for not doing enough to help its i ts citizens.
What happened in Osaka, Taiwan’s Taiwan’s Digital Minister Audrey Tang said, was a situation “where people collectively look back and start to think, maybe we rushed to conclusions. Maybe we should have stopped and said: ‘Is it true or not?’” Tang, who became Taiwan’s Taiwan’s first digital minister in 2016, rolled out the country’s pioneering vTaiwan vTaiwan digital platform pla tform for policy deliberation and debate which, by encouraging participatory behavior and open engagement with public officials, aims to help eradicate the climate of fear and distrust in which rumors and outright lies breed. Outside of government, Taiwan’s civic tech community, community, g0v (pronounced ( pronounced “gov-zero”), has produced tools such as Cofacts, a fact-checking social media chatbot, part of a culture of innovation in Taiwan that has flourished under the shadow of its neighbor, China. However, the threat of state-sponsored disinformation, which officials say is increasing as Taiwan’s Taiwan’s November 24 local elections approach, has led some top officials to discuss adopting a more direct strategy. strategy. Shortly after the September Kansai incident, the investigative bureau of Taiwan’s Ministry of Justice said it had found “unequivocal evidence” that Beijing was responsible for disinformation intended to divide Taiwanese society. society. The false Kansai Airport post, said Mainland Affairs Council (MAC) Deputy Minister Chen Ming-chi, “originated from a content farm in Shanghai.” “Our laws should be amended in response to the threat” posed by Chinese influence, said Chen, speaking at a conference held by the government-funded Institute for National Security and Democracy Research (INDSR) think tank on October 9. “Our laws should be able to prevent, prosecute, and punish any sabotage activities.” Taiwan is now considering amending its National Security Act to potentially criminalize the spread of false information online. Any eventual implementation of such considerations – which would require laws allowing the government to access citizen data – remains firmly within the realm of speculation. However, free speech advocates have reacted with w ith grave concern. Legislative approaches like those proposed by Taiwan “inevitably
end up becoming mechanisms for silencing dissent,” Ethan Zuckerman, the director of MIT’s Center for Civic Media, told The Diplomat at Diplomat at g0v’s October biennial summit. Taiwan is no stranger to propaganda: Throughout its long crossstrait détente with China, both the Republic of China (Taiwan) and People’s People’s Republic of China have blasted messages loud and soft at each other. The digital disinformation threat Taiwan now faces, however, is as new to the young democracy as it is to the rest of the world.
Taiwan’s Digital Minister Audrey Tang. Audrey Tang, Pixabay
When The Diplomat spoke Diplomat spoke with Tang in her office in Taiwan’s Executive Yuan, or cabinet, she had just attended a workshop held
by the Taiwan Foundation for Democracy at which Taiwan’s democratic defenses against disinformation received international praise. Taiwanese civil society takes immense pride in its democracy, democracy, but a question now lingers of whether its transparency serves as a strength, or as an Achilles heel. For Tang, Tang, openness is a natural defense against disinformation. She personally abides by a philosophy of “radical transparency,” transparency,” sharing written transcripts of everything from internal Cabinet meetings to interactions with lobbyists and journalists (including her interview with The Diplomat). Diplomat). This, she says, snuffs out rumors about her before they can spread. “If you have a good friend who you are meeting every week, if you hear gossip about that friend, of course you will ask” about it, she said. “But if that friend takes three months to get back to you… then of course there’s a lot of room for speculation.” A longtime civic tech pioneer and a nd hacker, Tang, 37, has gained notoriety for engaging online critics who post inflammatory messages – a practice she calls “troll hugging.” She said that, in her view, view, most faulty online information is i s spread by aggrieved netizens “who don’t get sufficient social interaction, like hugs or kisses, offline.” Tang distinguishes misinformation, misinformation, consisting of individual speculation from frustrated internet trolls along with misguided journalistic reports, from disinformation, disinformation, which she said consists of “organized and intentional” campaigns “ignoring “i gnoring context and basically sowing discord, fear, uncertainty, uncertainty, and doubt.” In Taiwan, both can be described by the Mandarin term jia term jia xinwen, or xinwen, or “fake news,” a term Tang refuses to use in either Mandarin or English. The term also leads to an incorrect conflation between false information with domestic origins, regardless of intentionality, intentionality, and organized foreign-sourced disinformation. Experts, such as Nick Monaco of Oxford University’s Computational Propaganda Project, agree that the overwhelming majority maj ority of Taiwan’s so-called jia so-called jia xinwen still comes from within the country. Tang prescribes media literacy as an antidote to immunize the populace against faulty information. Next year, new curricula will
be introduced to grades one, seven, and 10 that will teach children to critically analyze news reports and advertisements. Tang hopes this will “mold minds … [to] become less susceptible to the topdown broadcast messaging era.” She also spearheaded the Executive Yuan’s “Real-Time News Clarification” website, which aggregates government responses to information that may be false or dubious. Launched in May 2018, the website has posted 834 clarifications as of October 22 and generally provides public explanations within four to five hours of a news release. “When people see speculation or rumors in the morning news,” said Tang, “they get into the habit of waiting until noon” for a government clarification. Tang insists the cabinet clarification website is not a fact-checking tool. Its role has received criticism from skeptics such as Lin Fuyueh, director of Public Affairs at a t Taiwan Media Watch, a press freedom watchdog that, along with the Association for Quality Journalism, supports the independent Taiwan Taiwan Fact Checking Center. Lin told me he worries that the website labels items as misinformed when “the media or people disagree with the government’s government’s idea,” citing a recent item on a controversial coalfired power plant proposal. Tang, however, stressed that the website exists for the government to proactively provide its viewpoint. She does think the website can at times confuse its respective responses to misinformation, disinformation, and criminal offenses. To fix this, she said, it will soon roll out a color-coded or tagged hierarchy to allow users to sort by category. category. She also noted that opposition parties are, of course, free to set up their own platforms. “It’s great that [critics] can talk candidly,” candidly,” she said, pulling up a world map of press freedom by country on a projector and pointing to Taiwan, alone in a sea of relatively unfree countries. “We take pride in being the only green dot there.” The line between disinformation and criminality criminali ty,, however, is now at stake in Taiwan. Resulting legislation, if deemed to overreach, could potentially threaten Taiwan’s green hue on the press freedom map.
Prior to discussions of amending the National Security Act, legislator Chiu Chih-wei of the ruling Democratic Progressive Party (DPP) proposed an amendment to Article 63 of Taiwan’s Social Order Maintenance Act, which would make the spread of false information online punishable by fines or detentions of up to three days. The idea was slammed by opposition legislators, netizens, and the Committee to Protect Journalists. While momentum for such an amendment has petered out, some officials do insist that Taiwan use the “sword of the state,” as DPP legislator Cheng Pao-ching put it when blasting the National Communications Commission (NCC) for not using a law allowing it to punish broadcasters who fail to verify facts and promote fairness in news and commentary. commentary. At the October 4 hearing, NCC chairwoman Nicole Chan responded to harsh attacks from incensed fellow DPP legislators by saying free speech concerns had led to that particular sword being kept in its sheath. The following week, Chen, the MAC deputy minister, told an audience of ministerial officials, foreign envoys, and private sector defense representatives at the INDSR think tank conference – held to discuss responses to Chinese cybersecurity and “sharp power” threats – that Taiwan must embrace a punitive approach as it “navigated in the troubled water” between free speech and national sovereignty. “Given Taiwan’s Taiwan’s authoritarian past,” said Chen, “our civil society is highly concerned if the government takes any action that might constrain the freedom of the press.” In June, Chinese journalist Ye Qinglin was barred from reporting in Taiwan for spreading “false information through fake news” and “creating cross-strait conflict,” according to a MAC M AC statement made prior to Chen’s appointment as deputy minister. The decision, noted Chen, was widely wi dely criticized by Taiwanese press freedom advocates. But he said that Taiwan cannot treat Chinese media in the same manner as the United States; in September, the U.S. forced Xinhua forced Xinhua to to register as a foreign agent. Cracking down further on foreign peddlers of journalistic falsehoods, he said, would require public support that does not currently exist.
“The political will is simply not there,” said Chen. “We would like to see if the people become angrier with the media and give us some space to take some action.” Ma Ying-han, the head of the military’s cybersecurity command unit, said at the conference that, as Taiwan responds to an escalating Chinese campaign of cyberattacks and disinformation, “there’s an ethical line that we will watch over and not cross.” Taiwan, which endured decades of martial law under Chiang Kaishek and held its first democratic presidential election in 1996, is understandably wary of approaching such a line. It is now grappling with the question of when, exactly, exactly, false rumors and outright online lies transition from protected speech into a matter of national security. security. Tang is promoting a digital communications bill, spawned from participatory discussion on vTaiwan, which she said “links the penal code that's regulating the existing, offline world and say[s] these behaviors have their equivalence online.” The bill would thus provide a framework to apply existing laws against social discord (such as making harmful threats) to the online world. “Before, it was kind of a grey area,” she said. The law would also put into law a clear distinction between misinformed yet legal online speech and criminality. criminality. If disinformation leads to someone committing a criminal offense, she said, “of course it will be a threat to national security.” Potentially thorny questions as to which online speech should be defined as criminal in nature would be sorted out by democratic processes. “For any emergent issue, there should be a multistakeholder conversation with civil society, society, the social sector, and international actors” once the bill passes, said Tang. The bill is scheduled to be debated and voted upon by Taiwan’s Taiwan’s legislature this autumn. Wu Jun-deh, a researcher in INDSR’s INDSR’s division of cyber warfare w arfare and information security, security, told me Taiwan needed quicker, more effective methods to counter an improving Chinese cyber force.
In recent years, said Wu, meddling attempts would succumb to rookie mistakes such as using simplified Chinese characters (Taiwan uses traditional characters). He pointed to a May PTT post falsely claiming an envoy from Honduras, one of Taiwan’s Taiwan’s 17 remaining diplomatic allies, was visiting Beijing in anticipation of switching its recognition to the PRC. The story, story, which was quickly refuted, was sourced to a Chinese IP address after observers noticed its use of a term for “Honduras” used in China but not in Taiwan, he said. Wu did stress that sourcing information was difficult – echoing Tang, who noted that IP addresses, which are incredibly easy to spoof, cannot be relied upon. The Honduras PTT post, however, was identified by an anonymous national security official as being linked to a Chinese state-sponsored disinformation mill with ties to a task force responsible for spreading false information within Taiwan. In his remarks at the INDSR conference, Ma said the so-called 50 Cent Party – the infamous, shadowy group of Chinese internet commentators that allegedly publishes articles and forum posts on behalf of the Chinese government – is expanding its campaign of automated disinformation. However, much like Russian meddling in the 2016 U.S. presidential election, the exact scope and intent of such organized, bot-driven disinformation campaigns remains unclear. As analysts continue to try and quantify the threat, Lin of Taiwan Media Watch firmly believes that politicians in the ruling DPP should not be involved in regulating speech. “They worry about the security of our nation. They worry about [remaining] the ruling party of Taiwan now,” he said. “We want to send a message to government: Don’t do this. This will be very bad for free speech.”
Huang Chao-hui points to Taiwan on a map illustrating Reporters Without Border’s 2018 free press rankings. Nick Aspinwall
Lin, while not a fact-checker himself, is a regular presence at TFC’s humble offices. Tucked into a small third-floor walkup in the outskirts of Taipei, the center began operating in August, bringing on veteran journalist Huang Chao-hui to serve s erve as chief editor overseeing a small four-person team of fact-checkers. Nobody at TFC, just a few weeks into its existence, expected a disinformation whopper on the level of Kansai. The center quickly ran into financial issues during its response to the false Kansai story, said Huang, who is also an assistant professor at National Taiwan University’s journalism school. “We didn’t have the money to hire an expert translator,” she said. Eventually, Eventually, she enlisted one of her Japanese-speaking students to
translate a list of seven questions which were sent to a factchecking center in Japan and forwarded to the Japanese government. TFC was initially hesitant to enter the post-Kansai social media frenzy, frenzy, holding out hope that Taiwanese news outlets would find the truth. “We didn’t jump in immediately,” Huang told me. “We were hoping the media would come up and do some clarification. After three to four days, we realized nobody was doing that.” Huang echoed a common refrain of Taiwanese reporters when she said the country’s competitive media environment prioritizes immediacy over accuracy, creating a fertile ground ripe for exploitation by bad actors. “The responsibility is still on the media to fact-check, even though they have very limited time and energy,” energy,” she said. “After the Kansai case, the truth was revealed, but nobody saw it. The fake news kept spreading.” Her team of fact-checkers has workshopped solutions that reach beyond its limited scope – as of the time of writing, TFC has published 24 fact-check reports in just under three months, with at least three editors laying eyes on each item. Huang vehemently opposes a legislative approach to resisting false information, noting that TFC already faces pushback from some reporters it fact-checks who, eager to maintain their own relationships with politicians, recite incorrect or contradictory information provided by public officials. Instead, Huang wants the government to pressure social media companies to support media accuracy, accuracy, either via user experience (for example, clear demarcations on Google search results indicating i ndicating a story’s accuracy) or through funding fact-checkers in news outlets. Wu of INDSR believes this is the most prudent solution, arguing that Big Tech must take “social responsibility” even at the expense of profitability. profitability. Both Huang and Wu mentioned Germany’s Germany’s NetzDG law against internet hate speech, which requires social media companies to remove offensive posts within 24 hours but has in some cases driven more attention to the voices it attempted to muzzle.
“You “You don’t want to silence people,” said MIT’s Zuckerman, “and have them disappear and go underground to the dark web, to places you can’t interact with them.” “The tools to counter disinformation can be weaponized by bad actors,” warned Scott Hubli, director of governance programs at the Washington-based National Democratic Institute. In some instances, he told me at October’s g0v summit, “trust in government is such that government isn’t always the most credible actor to respond. Sometimes, an independent actor is far more credible.”
Johnson Liang (left), founder founder of Cofacts, explains explains the LINE fact-checking fact-checking chatbot's functionality functionality at g0v Summit 2018. g0v Summit 2018
One actor operating in this space in Taiwan is Cofacts, an opensource chatbot that allows users of the popular private messaging app LINE to submit links or messages for fact-checking by a team of volunteer editors. According to a 2016 Nielsen survey, survey, 91 percent of Taiwanese aged 12 to 65 use LINE every week. The app is also particularly attractive to disinformation merchants. According to Tang, “they know these are closed channels” where rumors can rapidly spread before appearing on open platforms like Facebook. Cofacts, which originated from a 2016 g0v hackathon, aims to penetrate the shadowy annals of LINE chat groups which, like those of WhatsApp, are rife with scams and falsehoods. Editors respond to about 250 inquiries each week from over 45,000 LINE accounts that have downloaded the chatbot; real-time data visualizations are publicly available online. Replies to popularly submitted messages, said Cofacts founder Johnson Liang in an open response to questions from The Diplomat, generally take one to two days to turn around. The chatbot responds to about 70 percent of queries, according to Liang, and judges just over a third of them to be false. “In LINE chatrooms I am in, I would receive internet hoax[es] and rumors from time to time,” said Liang. “Cofacts makes netizens like me less stressed when faced with such hoax[es].” As electoral politics have taken hold of Taiwan, Cofacts has fielded numerous inquiries on various candidates featuring “persuasive opinion and propaganda” as of late, said editor Billion Lee. Items on LGBT issues are also frequently submitted – Taiwan is scheduled to hold referendums on same-sex marriage and sexual education in November – as are rumors on public health, said Liang. Cofacts, however, is not designed at present to counter an “internet army” like China’s 50 Cent Party – nor can it insulate itself from potential infiltration such as what plagued PTT, PTT, which suspended new user applications after the Kansai Airport incident due to what a system administrator called a “loose regulation process.”
“We are as vulnerable as PTT is, since we do allow all to become editors,” said Liang. He hopes that, like PTT, PTT, Cofacts will be able to employ “creative counter measures” to root out mis- and disinformation players as it grows in scale. At present, Cofacts stands as one of the creative, laudable civic lines of defense rolled out by Taiwan against the threat of crossstrait online falsehoods darting across the screens of Taiwan’s Taiwan’s social media-crazed population. “The war of the 21st century,” century,” said Huang of TFC, “is information.” Taiwan is venturing into this present-day digital battlefield facing the same hefty civic dilemmas as many of its democratic counterparts. Its response may dictate whether it can be, as U.S. Deputy Assistant Secretary of State Scott Busby said in remarks at the Taiwan Foundation for Democracy, Democracy, an “invaluable model” for its Asia-Pacific neighbors.
The Author Nick Aspinwall is a freelance journalist based in Taipei.
