China’s meddling in Taiwan election opens year of misinformation threats - The Washington Post

notion image
KMT presidential candidate Hou Yu-ih greets supporters during a motorcade campaign tour in Taipei, Jan. 9, 2024. (I-Hwa Cheng/AFP/Getty Images)
A sophisticated Chinese Communist Party effort to tip Saturday’s election in Taiwan may establish a template for interfering elsewhere ahead of a wave of critical global elections, analysts in multiple countries said.
The close contest between the ruling Democratic Progressive Party (DPP) and the more mainland-friendly Kuomintang(KMT) is the first major vote in a year that also features electoral battles in Indonesia, Pakistan, Mexico, Russia and the United States.
In addition, the European Union will host elections for its Parliament, which could affect the direction that the 27-member bloc takes on key policies such as migration, while British Prime Minister Rishi Sunak has said he will call elections later this year, as the country’s ruling Conservative Party grapples with an economy tipping toward a recession.
And India will hold nationwide elections for the lower house of Parliament, a test for Prime Minister Narendra Modi and his Hindu nationalist Bharatiya Janata Party. A Washington Post investigation last year revealed how the BJP and affiliated nationalist groups have used social media to cement their grip on the electorate.
In total, more than half the world’s population lives in countries holding elections in 2024 — creating an unprecedented test of the systems that governments, tech companies and researchers have built to defend democracies from disinformation.
While the Chinese government is the biggest power interfering in Taiwan, which Beijing views as part of its territory, it is just one of multiple nations likely to meddle in elections elsewhere this year, especially in the United States, experts said. China, for example, has been ramping up U.S. information operations and triggered multiple takedowns of fake accounts at Facebook, while Russia is trying to dissuade European countries from supporting Ukraine.
The rise in election interference comes as tools for disguising where messages originate are getting better, while the major social media platforms are cutting back on rules and enforcement. Meanwhile, few nations called out for interference in the past have been punished, beyond sanctions against some Russian officials and executives.
The Taiwan election is “the canary in the coal mine,” said Katie Harbath, a former public policy official at Meta, owner of Facebook and Instagram. It’s “a sense of what we might see throughout the rest of this year.”

Propaganda war

Taiwan has long been a proving ground for mainland propaganda campaigns. Some researchers have described it as the most contested information space in the world, against the backdrop of worsening military tensions.
The Chinese Communist Party (CCP) sees its work on propaganda aimed at the islanders as a core responsibility — one that ordinary Chinese citizens, including those who have emigrated, may be called upon to help carry out, according to Anne-Marie Brady, a professor and Taiwan specialist at New Zealand’s University of Canterbury.
At the same time, the CCP has been less direct as the election approaches, wary of overdoing it, experts note.
“There’s a real potential for it to backfire,” said Alexander Dukalskis, author of “Making the World Safe for Dictatorship.”“People don’t want to be bullied and intimidated into being told how to vote.”
Instead, the Chinese government has used proxies, such as Taiwan’s business elite who earn money from trade with the mainland, inviting them on subsidized tours of their large neighbor. Local Taiwanese officials, now under investigation by local prosecutors, are also targets.
And rather than push their own messages, the propagandists have been encouraged to amplify authentic local disputes and divisions, said Tim Niven, head of research at Taiwan’s Doublethink Lab.
Propagandists have also been quick followers of local news, putting together clips from the most incendiary comments on talk shows and giving misleading summaries.
Generative artificial intelligence and other new tools are helping, Niven said.
Fake news videos, with AI-generated hosts and voice-overs, have circulated on YouTube, Instagram and X, according to a Taiwanese national security official’s accounts to local media in Taipei.
Representatives for Meta and X did not immediately respond to requests for comment late Thursday.
YouTube spokesman Javier Hernandez said the company has removed a number of channels violating its content rules ahead of the Taiwan vote.
“We have teams dedicated to combating coordinated influence operations and are working around the clock as we approach the Taiwanese elections,” he said.
Just last year, such attempts were far less sophisticated. But the technology has gotten much better in a short time, notably in the ability to create AI-generated images or clone voices, said analyst Libby Lange of the misinformation tracking company Graphika.
“It’s just such a leap forward in scary ways from where we were before,” Lange said. “If everything could be fake … there’s really no sense of ground truth.”

Multipronged attack

All of those tactics — the use of AI and local allies, issues and news — are hard to combat, and all can be replicated around the globe.
In a campaign exposed by Graphika last month, an account known on various social media platforms as Agitate Taiwan posted short videos on TikTok and YouTube that criticized candidates and policies of parties other than the KMT, which antagonizes the mainland less. Those videos were simultaneously posted by some 800 Facebook accounts to various Taiwanese Facebook groups, including those devoted to nonpolitical topics.
Graphika could not determine whether the Agitate Taiwan accounts on YouTube and TikTok were part of the campaign or if they belonged to a genuine user whose content was repurposed by the operation. The YouTube channel has been suspended, according to Graphika, but TikTok left the account up because it could not determine whether it was inauthentic, according to a company representative, who spoke on the condition of anonymity because of the risk of retaliation from parties spreading disinformation.
In another example, a spurious biography of Taiwan’s current president has gone everywhere in a blitz over the past few days, noted Lange. “The Secret History of Tsai Ing-wen” was sent by WeChat accounts in direct messages, and its content has appeared on every platform, according to analyst Ari Ben Am of Telemetry Data Labs. Some accounts on Facebook posted an electronic version of the book over and over.
The use of multiple platforms means that an effective defense needs coordination across those platforms and roughly similar policies. But the number of platforms used in the United States and other countries has increased, while X and Facebook have lost dominance. Foreign agents and homegrown conspiracy theorists often start on lightly moderated platforms such at 4chan and Telegram, sowing seeds that viewers can replant on more mainstream platforms.
Representatives for 4chan and Telegram did not immediately respond to requests for comment late Thursday.
For their part, Meta and Google have reversed policies against election-rigging falsehoods and stopped punishing the politicians who spread them. X has reinstated a slew of far-right accounts, gutted a program designed to verify high-profile users and failed to stop an influx of hate speech and misinformation.
Coordination is more challenging than ever, especially connecting with the diminished enforcement staff at X, according to a half-dozen people who teamed with them in the past.
The biggest problem in Taiwan is the one that affects the political landscape in the United States and elsewhere: The issue of foreign interference has become politicized. The KMT has tended to play down the severity of the problem, while the ruling Democratic Progressive Party has continued to point it out.
“Being able to talk about interference when the CCP is demonizing the DPP and promoting the KMT, it gets politicized,” said New Zealand’s Brady.
In the United States, a similar story is playing out as many Republicans reject any accusation of Russian interference, even after federal indictments of Russian spies and contractors who operated networks of bogus social media accounts.
Even when Chinese fakes are betrayed by technical means, “whether you believe it or not depends on your leanings,” Niven said.
Analyzing more than 10,000 YouTube videos on a few channels since June, Johns Hopkins School of Advanced International Studies researcher Martin Wendiggensen found that “the DPP is almost always described negatively and mentioned in negative contexts (corruption, incompetence, stagnation) while the reverse is true for the KMT.”
The channels are owned by wealthy Taiwanese business executives who generate most of their money on the mainland, Wendiggensen said.
Brady and other analysts said that the local government and civic fact-checking groups have been doing a good job calling out the interference as much as possible and educating the public. But CCP efforts could make the difference in the close election, they caution.
As with Russia’s approach in past U.S. elections, the main message — that the DPP is recklessly pursuing independence and risking war — does not have to convince a majority to succeed. Rather, it just has to cause division, “uncertainty and anxiety in society,” Brady said. “When people are afraid, they don’t want to take risks.”
Moreover, Taiwan is not like the United States, where tech giants are “going to have all the pressure in the world to get [U.S. elections] right,” said Rose Jackson, the director of the Democracy and Tech Initiative at the Atlantic Council. “They do not have pressure to invest the resources to get the rest of the world right.”
Even in the United States, those companies can’t catch all the coordinated, inauthentic surges of propaganda, and the fallbacks are getting harder to find.
For example, some social networks have said that the U.S. government has stopped warning them about foreign disinformation campaigns on their platforms, reversing a years-long strategy to fight international meddling in American politics.
The shift came as a federal judge limited the Biden administration’s communications with tech platforms, following a lawsuit that alleged that the White House’s coordination with industry to remove falsehoods about covid-19 and 2020 election amounted to unlawful censorship. That case, Missouri v. Biden, is now before the U.S. Supreme Court, which has paused lower court restrictions while it reviews the matter.
Meanwhile, disinformation researchers are reevaluating their efforts to track online falsehoods and alert tech companies amid an ongoing investigation by House Republicans, who are demanding documents and testimony from scholars about their interactions with tech companies and government, accusing them of colluding with the Biden administration to stifle American users’ voices online.
“It can both be true that companies are doing a lot to prepare, and it can also be true that just the sheer number of elections happening this year was going to test even the best-resourced of teams,” said Harbath, who is now the global affairs officer at Duco, a technology consulting firm. “Everyone’s trying to navigate all of these new variables that didn’t exist.”