China's AI-driven election meddling in Taiwan points to 2024 risks in US

Taiwanese groups on Monday outlined a major disinformation campaign employed by Chinese actors during the island nation’s national elections, which concluded in January and delivered a blow to Beijing with the victory of a pro-U.S. candidate.  

The alleged tactics from Chinese actors involved using generative artificial intelligence (AI) to manipulate videos and sow discord in Taiwan, including by completely distorting the words of at least one U.S. member of Congress, according to Taiwanese officials and nongovernment groups who briefed reporters at the National Press Club.

Many of the disinformation narratives focused on the U.S., such as falsely accusing Washington of building biological labs in Taiwan or fomenting the war between Israel and Palestinian militant group Hamas in Gaza.

The narratives harbor clues as to how China is trying to manipulate public opinion during the U.S. presidential elections this year.

Other Chinese disinformation narratives found by the Taiwanese groups refer to the U.S. as a fake ally that will abandon Taiwan, accuse Washington of having a fake democracy and blame Americans for sowing chaos across the world — all while Chinese propaganda portrays Beijing as the source of good and order.  

Chihhao Yu, the co-director of the Taiwan Information Environment Research Center, a nongovernmental organization (NGO) specializing in information research, said his organization has documented 84 narratives propagated by China seeking to stoke U.S. skepticism.  

Chihhao said Chinese actors are pushing to “fortify and reinforce” a pro-Beijing “alternative worldview,” which he said his NGO is combating by data collection and focus group discussions on the local level to understand the impact of information manipulation. 

“We need to figure out how that is reaching from our phones to our minds and how that is changing our perception of the world and how that is damaging our shared reality,” he said of the spread of disinformation. "We no longer share a shared reality."

The campaign in Taiwan is especially concerning for Washington because China is using similar tactics in the U.S., according to Microsoft, which released a report this month on Beijing’s disinformation campaigns. 

Some Chinese government-linked accounts have posted about presidential candidates in the U.S. elections, Microsoft said. One of the most prolific Chinese actors spreading propaganda is commonly known as Spamouflage, and another is an espionage group called Gingham Typhoon, which is very active in the Indo-Pacific. 

Microsoft warned that Chinese actors “honed their techniques and experimented with new media” in the Indo-Pacific last year, using a complex strategy of AI-generated news anchors and memes, and may improve their ability to influence elections down the line. 

“We are prepared to see influence actors interact with Americans for engagement and to potentially research perspectives on US politics,” Microsoft warned in the report. “China will, at a minimum, create and amplify AI-generated content that benefits their positions in these high-profile elections.” 

Russel Hsiao, the executive director of the nonprofit research and Washington-based Global Taiwan Institute, said China is “undermining the credibility and reliability of the United States” in Taiwan and other countries. 

“So all the more important that there's stronger and better and more efficient public-private partnerships within Taiwan, between the civil society and the government, but also internationally between governments and also civil society as well,” he said. 

The NGOs said the disinformation campaign in Taiwan was spread on popular apps, such as video-sharing platform TikTok, which in the U.S. is being hotly debated. Some lawmakers want to ban Chinese-owned TikTok as a national security threat. The House passed a bill in March that would ban TikTok unless it is sold by its Chinese parent company, ByteDance.

Analysts are increasingly warning that China is seeking to influence the U.S. elections this year. 

The Center for American Progress, a U.S. think tank, warned in a report earlier this year that China’s AI-driven disinformation efforts “have intensified.” 

Researchers wrote that Chinese actors are attempting to acquire Taiwan social media accounts and to “pay influencers to promote its narratives.” 

“It uses these methods to inundate online spaces with high volumes of content to manipulate algorithms and increase spread,” they wrote. “As the United States prepares for 2024 elections, Washington can examine both China’s disinformation influence efforts in Taiwan and the effectiveness of Taiwan’s countermeasures.” 

The Taiwanese elections were held in mid-January. The victor, Lai Ching-te, whose pro-U.S. party currently governs Taiwan and rejects Chinese sovereignty over the island, beat out a candidate from the more pro-China Kuomintang party. 

The election was closely watched by both the U.S. and China as tensions grow over a potential Chinese invasion or blockade of Taiwan. The U.S. has informal relations with Taiwan but commits to supporting Taipei. Chinese leader Xi Jinping has told his troops to be ready for a potential invasion by 2027. 

During the elections, along with the disinformation campaign, China also intimidated Taiwan over the strait separating the two nations by repeatedly flying aircraft and spy balloons across Taipei’s airspace and exclusive zone. 

The election was ultimately a setback to Beijing as it solidified Taiwan’s preference for closer ties with the U.S., and it showed that voters were not largely influenced by Chinese propaganda. 

Still, Eve Chiu, the editor-in-chief of the NGO Taiwan FactCheck Center, described a wide-reaching Chinese disinformation campaign that included narratives about the U.S. as the originator of COVID-19 and that the U.S. would fail to protect Taiwan in a war. 

Eve said the scope of the disinformation campaign was bigger than the last election in Taiwan, with AI boosting the spread of the content.  

China used generative AI to spread conspiracies about the CIA interfering in the elections and to manipulate the words of Rep. Rob Wittman (R-Va.) to say the U.S. was accelerating arms deliveries to Taiwan, she added. 

“It's very hard to verify,” she said, saying the videos are “very sophisticated.”