Taiwan’s Election: 2024’s Canary in the Coal Mine for Disinformation against Democracy

December 19, 2023

As we step into 2024, an unprecedented four billion people globally are gearing up to participate in what is set to be the largest election year in history. At the forefront of this democratic marathon is Taiwan, whose January elections not only inaugurate the year’s electoral calendar but also serve as a critical lens into China’s evolving prowess in information manipulation and electoral interference. Historically, Chinese information operations have been perceived as relatively ineffectual in the Anglophone world. However, China’s shared linguistic and cultural ties with Taiwan have made the island an ideal proving ground for China to hone both its information and malign influence strategies. Here, China has adeptly learned to deploy bots, mimic or recruit local influencers, and cross-leverage different sources of diplomatic and commercial strength to amplify its impact.

Next year will also witness the burgeoning role of artificial intelligence in shaping electoral narratives. A glimpse of this was seen in Argentina’s recent election, where AI-crafted content played a significant role in major political campaigns. This global trend will play out prominently in Taiwan, which is experiencing a surge in AI-generated disinformation, including images, videos, audio, and text, all aimed at influencing electoral outcomes. The island not only finds itself at the epicenter of this new wave of tech-enabled electoral meddling but is also pioneering responses to navigate this unprecedented challenge in digital democracy. As the world navigates a year of critical elections, attention should be keenly focused on Taiwan’s experience.

Beijing’s Objectives and Tactics

Beijing has been deploying information warfare through social media in Taiwan since at least 2016. That year, Taiwan elected President Tsai Ing-wen of the Democratic Progressive Party’s (DPP). Taiwan’s founding party, the Kuomintang (KMT), made an agreement with the Communist Party of China (CCP) that there was only “One China”. Tsai and her party maintain that Taiwan is “a sovereign, independent country called the Republic of China” (ROC), and does not need to risk the peaceful status quo by declaring formal independence from the People’s Republic of China (PRC). In short, they rejected the KMT’s so-called 1992 Consensus. Her election signaled a fundamental shift in Taiwan, from older generations who thought of themselves as primarily Chinese, to younger generations who thought of themselves as primarily Taiwanese. Beijing sees that trend as a ticking timebomb that will thwart their hopes of peaceful reunification with the island.

The PRC’s early influence campaigns in 2016—as well as in the years after—tried to undermine Tsai’s support in three ways: first, through pro-unification messages meant to enhance China’s image with potential pro-unification voters; second, through clumsy disinformation tactics, posting from official government accounts, and pushing obviously false content, like falsely claiming Tsai’s doctorate from the London School of Economics was a fake; and lastly, by trying to undermine the US-Taiwan relationship by portraying the United States as stoking war with China. In 2018, Chinese news web site Xilu (西陆) seeded a false story that Taiwan was planning to controversially lease Taiping Island to the United States for military use.

Improved Methods and Messaging

President Tsai’s administration withstood eight years of these influence campaigns, and in that time, China’s operations have become larger, more sophisticated, and more multi-modal. These improved methods include traditional approaches like spreading information from dedicated content farms through bots and fake accounts. Meta recently reported that it broke up the largest-ever such Chinese influence campaign on its platforms.

China’s disinformation peddlers have become more sophisticated in their messaging, which is less focused on cross-strait relations, and more focused on existing partisan and generational divides in Taiwanese society. A sampling of disinformation spread on LINE, a popular messaging app, in the lead-up to the 2022 referendum to lower the voting age to 18 showed several PRC-aligned messages aimed at mobilizing older “No” voters by telling them the threshold for passing the motion was 25% of the vote, when it was actually 50%. The motion, which would have enfranchised younger voters who tend to be more anti-PRC, failed to pass.

Over time, Chinese propagandists have become more adept at boosting organic Taiwanese content that is either friendly to Chinese interests, or deepens existing divisions in Taiwanese society. When propagandists can’t find the right message organically, Chinese officials have paid local social media influencers and marketing companies to spread pro-PRC content. Outside of social media, China’s United Front has penetrated Taiwan’s business and media communities after decades of engagement. Due to a lack of transparency over foreign funding in media, lobbying, politics, and business, it can sometimes be difficult to ascertain what is Chinese influence and what is sincerely held opinion.

Beijing has also refined its strategy for tarnishing the United States’ reputation as a partner for Taiwan. Online, it uses all of the techniques described above to boost existing “US Skepticism” narratives, which claim the United States takes advantage of Taiwan like a colony, uses Taiwan to create bioweapons, and colludes with corrupt DPP officials to take advantage of the Taiwanese people. Offline, China uses cognitive warfare by way of constant military intimidation to persuade Taiwanese voters that the United States and pro-independence parties will provoke a conflict. In recent years the Chinese military has carried out live fire military drills around the island, sailed Chinese navy vessels into Taiwanese waters, flown war planes through Taiwanese air space, and have flown combat drones around Taiwan’s coasts. Chinese influence operators often pair these hostile maneuvers, like those in retaliation to Speaker Nancy Pelosi’s visit to Taiwan in August 2022, with a barrage of propaganda across multiple channels for maximum effect.

Generative AI and Disinformation

Perhaps the most worrying new strategy for influencing Taiwan’s upcoming election is using generative AI to create images, audio and videos convincing enough to cause uncertainty. Earlier this year, videos of outgoing President Tsai, and her Vice President Lai Ching-te, the DPP’s candidate in the 2024 elections, circulated on social media, showing them promoting investments in cryptocurrencies. The videos, accompanied by fraudulent links to “invest”, used deepfake technology to mimic their voices, and adjusted original footage to make their lips match their words.

In another case, an audio clip was circulated showing the populist Taiwan People’s Party (TPP) Presidential candidate Ko Wen-je criticizing Vice President Lai, and accusing him of paying people in attendance at his campaign events. In both cases, the relevant Taiwanese authorities determined the content to be fake after an investigation. Incidents like these are almost certainly designed to tilt the election against the DPP, which takes a relatively harder line against the Chinese Communist Party, and to undermine trust in both the election and media coverage surrounding it.

Although there’s no concrete evidence yet linking the PRC to these particular AI-generated fakes, the content’s alignment with Chinese interests strongly suggests official state involvement. Beyond speculation, there’s substantial evidence that China has been developing these capabilities. The PLA’s Strategic Support Force, particularly the Base 311 unit, focuses on political and psychological warfare and has discussed creating AI-generated disinformation, including deepfakes and tailored content. An article published in 2018 by Base 311, in collaboration with the National University of Defense Technology, details strategies for using advanced technologies like deepfake and big data analytics for propaganda purposes. These strategies are designed to blur the origins of such content, underlining the intricate challenges faced by those attempting to trace and counteract these sophisticated disinformation efforts.

Democratic Resilience

It is important to emphasize that while Chinese influence operations are a serious threat, Taiwan’s democracy is not a fragile target. Partially because of the relentless barrage of disinformation, the country’s model for countering disinformation is one of the world’s most developed. The island’s Disinformation Coordination Team (DCT) leads a coordinated interagency approach to define, monitor, combat, and debunk disinformation. Taiwan also has an independent ecosystem of independent civil society organizations which provide toolsresearch, and fact-checking aimed at combating disinformation.

The advantage of this model is that it works along two parallel tracks which complement one another. Civil society groups identify, and ideally can make attributions for, disinformation. The government, through bodies like the DCT, can draw on these findings and respond by debunking claims at scale by providing relevant, timely information to the public. Taiwan should explore options to expand funding for civil society groups tackling this issue without undercutting the editorial independence that makes them so trusted. Taiwan should also encourage the adoption of crowdsourcing techniques that augment the capacity for detecting and reporting potential disinformation at low cost through volunteers, especially when used in combination with AI-enabled counter-disinformation tools.

However, neither civil society groups nor governments can tackle the deluge of disinformation alone, even in battle-hardened Taiwan, where US social media giants have a strong presence. Meta played a positive role in combating disinformation in Taiwan’s 2020 election, and its recent takedown of Chinese campaigns is commendable. However, content remains up on other platforms, and there is a pressing need for technology companies to reinvest in their disinformation capabilities. Meta, Google, and X (formerly Twitter) have all massively cut disinformation staff, leaving, for example, only one person in charge of misinformation policy worldwide at YouTube. X has curtailed access to data needed to study disinformation on its platform. Taiwan should continue encouraging these technology companies to reinvest in their capabilities, which would greatly expand the island’s capacity to combat disinformation.

Many of these measures will take months, if not years, to reach their full potential. In the meantime, Taiwan’s election is only a month away, and short-term measures will be crucial to mitigate the influence of disinformation campaigns and ensure the integrity of the upcoming election. The Taiwanese government should continue to ramp up its rapid-response teams to respond to disinformation, and communicate proactively about efforts to safeguard the election.

What this Election Means for Taiwan and the World

For Taiwanese voters, this election is primarily about a host of domestic issues. Lai Ching-te’s task is to convince the electorate to give the DPP an unprecedented third term in a row, but voters of all stripes have grown frustrated with the economy, and prices that have grown faster than their wages. Many young DPP supporters have grown frustrated with a lack of movement on progressive priorities, and have moved towards the TPP’s Ko Wen-je, who has a reputation as being a populist reformer focused on social justice, though his brash style has led to controversies and accusations of sexism.

Meanwhile, the rise of the TPP has put the KMT on the backfoot. They have struggled to defend their status as one of Taiwan’s two major parties. The KMT’s candidate, Hou Yu-ih, has made healthcare policy a major domestic focus, but on foreign policy Hou has argued that the KMT is the party of peace, whereas a vote for the DPP would lead to war. Given the KMT’s historical and contemporary ties to the PRC, this position is not surprising, but outsider observers should be hesitant to equate a vote for the KMT with a vote for surrender to China. While all three candidates support some version of the status quo arrangement with the PRC, they primarily differ on whether deterrence or engagement is the best tool to avoid war.

Nevertheless, the CCP’s preferred outcome would be a defeat of the DPP, whose policies over the last eight years have brought it closer to the United States. Beijing’s belief is that the other two parties would be easier to work with, especially the KMT, whose willingness to economically engage with the mainland could complicate Taiwan’s efforts to build resilience against coercion. Chinese influence operations are strategically aimed at swaying Taiwan’s election to favor an agenda of so-called peaceful reunification, or, in a more dire scenario, to render Taiwan a more vulnerable target for invasion.

Looking ahead, the PRC will advance its interests through influence campaigns not only in Taiwan, but around the world. The urgency for robust, immediate counter-disinformation measures is paramount, not only for Taiwan, but for democracies worldwide. This moment is about more than safeguarding a single election; it is about the broader battle democracies face in our evolving digital age. Taiwan’s efforts to fortify its own electoral integrity, and the lessons it will learn in the coming weeks, should provide a blueprint for all the year’s upcoming elections, especially the United States, as it prepares for its presidential election in November.

 

The views expressed in GMF publications and commentary are the views of the author alone.