AI War: Taiwan Experts Warn Against Chinese Election Disinformation
Taiwan academics urge the government to deploy AI filters and media literacy to fight Beijing's generative AI disinformation campaigns.
As local elections approach, academics and experts in Taiwan are sounding the alarm over a sophisticated surge in Chinese election disinformation. With Beijing increasingly leveraging low-cost generative AI tools to manipulate public perception, experts are urging the Taiwanese government to implement robust AI content filtering systems and aggressively promote media literacy among the citizenry.
The warnings came during a forum hosted by the Democratic Progressive Party (DPP) focused on Chinese interference. The gathering brought together specialists to dissect the evolving tactics used by Beijing to sway political outcomes and destabilize the democratic process on the island.
The '30/70' Strategy of Deception
DPP Secretary-General Hsu Kuo-yung highlighted a particularly insidious method of manipulation. He noted that China is not relying on outright lies, but rather a blend of truth and fiction, circulating information that is roughly "30 per cent true and 70 per cent false." This strategy is designed to make disinformation more believable while steering the narrative toward Beijing's goals.
According to Hsu, the integration of big data and AI allows Beijing to shape public perceptions with surgical precision, making it imperative for experts to analyze how these technologies are being weaponized.
Short-Term Wins, Long-Term Chaos
Yeh Yao-yuan, a professor and chair of international studies at the University of St Thomas in Houston, provided a broader perspective on Beijing's objectives. While the immediate goal is to influence the results of upcoming elections, the long-term strategy is far more damaging: deepening social fractures and eroding the public's fundamental trust in democracy.
Yeh explained that Beijing is employing a multi-pronged approach that includes:
- Public Opinion Manipulation: Using bots and coordinated accounts to shift narratives.
- Social Network Infiltration: Embedding influence agents within digital communities.
- Military Pressure: Using physical threats to create an atmosphere of instability.
Adding a chilling dimension to the threat, Yeh revealed that the US government discovered in July that China had stolen 220 million US voter registration records. He warned that if similar data analysis is applied in Taiwan, Beijing could identify hyper-local issues and tailor disinformation campaigns to target specific groups of voters with alarming accuracy.
The Rise of AI-Generated Deceit
The scale of the problem has been amplified by the accessibility of AI. Liu Yu-hsi, a communications professor at Shih Hsin University, pointed out that Beijing is now using AI-generated text and images to produce disinformation at an industrial scale.
Liu cited a telling example from this year: Chinese immigrants running for borough chief positions. Many of their online campaign graphics were riddled with errors, a hallmark of poorly executed AI-generated content, yet these materials are still used to penetrate the local political discourse.
The Evolution of Interference
Arthur Wang, secretary-general of the Asia-Pacific Elite Interchange Association, noted that Beijing's tactics have evolved significantly over the decades. He mapped the progression as follows:
- Early Phase: Reliance on verbal attacks and military threats.
- 2008 Era: A shift toward economic inducements.
- 2016–2018: An increased emphasis on general disinformation.
- Current Phase: Targeted election interference designed to intensify internal antagonism and undermine Taiwan's sovereignty.
The Roadmap for Defense
To fight back, Wang urged the government to develop specialized AI databases and content filtering systems. However, technology alone is not the answer; he emphasized the need for expanded media literacy education so that citizens can identify and verify suspicious content themselves.
Adding to this, DPP China Affairs Director Wu Chun-chih advised the public to be wary of information from unknown sources, especially content designed to trigger strong emotional reactions. He recommended a simple but effective defense: always check original sources and cross-reference claims with reputable, established media outlets.