In the quiet corridors of academic institutions and party offices across Taipei, a new kind of election battle is taking shape—one fought not with missiles or naval maneuvers, but with algorithmically generated images, half-truths, and digital whispers designed to sow chaos. As Taiwan braces for local elections, academics and political leaders have raised a collective alarm: Beijing is increasingly using cheap, powerful generative AI tools to flood the island with disinformation, blurring the line between what is real and what is fabricated. At a recent forum hosted by Taiwan’s Democratic Progressive Party (DPP), experts gathered not just to analyze the threat, but to offer practical solutions rooted in education, technology, and public vigilance. DPP Secretary-General Hsu Kuo-yung framed the problem in stark, memorable terms: China, he said, is circulating information that is “30 per cent true and 70 per cent false.” That single phrase captures the insidious nature of modern influence operations—they are not pure lies, which would be easier to spot, but carefully engineered distortions wrapped around kernels of truth. The goal is not simply to persuade people to vote a certain way, but to erode the very concept of objective reality in the minds of voters. In such an environment, a forged campaign poster, a manipulated quote, or a fabricated news story can travel across social media in minutes, lodging itself in the public consciousness before any correction can catch up.
Yeh Yao-yuan, a professor and chair of international studies at the University of St Thomas in Houston, offered a deeper framework for understanding what China hopes to achieve. He distinguished between Beijing’s short-term and long-term objectives. In the short term, the goal is straightforward: influence the outcome of specific elections. But the long-term goal, he argued, is far more corrosive—to deepen social divisions, widen existing political fractures, and systematically undermine public trust in democratic institutions themselves. This is not a new strategy, Yeh noted, but it has been supercharged by technology. Beijing’s toolkit now includes disinformation campaigns, public opinion manipulation, social network infiltration, and psychological pressure, all layered on top of traditional military intimidation. The result is an information ecosystem in which Taiwanese voters must constantly question not just the news they consume, but the authenticity of images, the source of viral posts, and the motives of anonymous accounts. Yeh also raised a deeply alarming claim: that the US government discovered in July that China had stolen 220 million American voter registration records. If true, the implications are global. With such data, he warned, Beijing could identify region-specific concerns of particular voter groups—whether it be housing costs, energy prices, or cultural identity—and tailor influence campaigns with surgical precision. Taiwan, he argued, must build stronger safeguards before the same methods are directed more intensely at its own electorate.
The role of generative AI in scaling these tactics was the focus of Liu Yu-hsi, a communications professor at Shih Hsin University. Liu pointed to a curious and telling phenomenon: this year, a number of Chinese immigrants had registered as candidates for borough chief positions in Taiwan, a low-level but symbolically important layer of local governance. Their online campaign materials, he said, were riddled with errors—incorrect addresses, strange fonts, impossible backgrounds, and other telltale signs of AI-generated text and images. In previous election cycles, producing disinformation at such scale would have required teams of writers, graphic designers, and translators. Today, a single operator with access to a generative AI tool can produce hundreds of convincing-looking posts, flyers, and even deepfake videos in a matter of hours. The cost is negligible; the potential impact is immense. What makes this particularly dangerous, Liu explained, is that AI-generated content does not need to be perfect. It simply needs to be believable enough to be shared, and frequent enough to normalize confusion. When voters are exposed to a steady stream of fabricated material, they may begin to dismiss everything they see online—including legitimate reporting and official announcements—as equally suspect. This is exactly the outcome that influence operations are designed to achieve: a paralyzed, skeptical public that no longer knows whom to trust. The response, Liu suggested, must be as innovative as the threat. Public education must evolve to teach citizens how to identify the fingerprints of AI content—subtle inconsistencies in hands, text, shadows, and voice—while platforms and government agencies must build robust databases of known disinformation patterns for rapid identification and removal.
Arthur Wang, secretary-general of the Asia-Pacific Elite Interchange Association, placed the current crisis in historical perspective, noting that Beijing’s tactics have evolved significantly over the years. In earlier decades, China relied on verbal attacks and military threats—a blunt approach that often had the unintended effect of rallying public opinion in Taiwan against Beijing. In 2008, Wang observed, the strategy shifted toward economic inducements, with promises of trade benefits and cross-strait business opportunities designed to create dependency. But between 2016 and 2018, the emphasis moved decisively toward disinformation, as China recognized the power of social media to shape perceptions without the need for direct confrontation. Today, Wang argued, Beijing’s approach is even more sophisticated: targeted election interference intended to intensify divisions and antagonism within Taiwanese society. This means amplifying tensions between political parties, ethnic groups, generations, and urban versus rural communities. It means engaging with online movements not to create them, but to fan their flames. It means using AI to identify which messages will most enrage or polarize specific communities, and then deploying those messages with mathematical precision. The broader objective, Wang said, remains unmistakable: to undermine Taiwan’s sovereignty by making its democracy look chaotic, dysfunctional, and unworthy of international support. If Taiwan’s elections become associated with endless conspiracy theories, bitter infighting, and public exhaustion, the story itself becomes the weapon. Wang urged the government to respond by developing AI databases and content filtering systems, and by expanding media literacy education across all age groups. People need to understand, he said, how AI can be used to create misleading information—and they need practical skills to identify and verify suspicious content before sharing it with friends, family, and neighbors.
On the practical side, DPP China Affairs Director Wu Chun-chih offered a simple but crucial piece of advice that resonates with anyone who has ever scrolled through social media at midnight and felt anger rise at a sensational headline: be cautious. Wu advised people encountering information from unknown sources, or content deliberately designed to provoke strong emotional reactions, to pause before reacting. Strong emotions—whether outrage, fear, or excitement—are exactly what disinformation engineers are trying to trigger, because emotional responses bypass critical thinking. His recommendations were refreshingly straightforward: check the original source of the information, look for official references, and cross-reference claims with reputable media outlets that have established track records of accuracy. These are not technical fixes requiring advanced degrees; they are simple habits that can be cultivated by anyone. But Wu’s advice also highlights a deeper challenge. In an age of information abundance, verification takes time and effort. It is easier to share an alarming post than to investigate its origins. Countering disinformation, therefore, requires a cultural shift—one that celebrates patience over impulse, curiosity over confirmation bias, and humility over certainty. The forum’s recommendations, as reported by the Taipei Times, reflect a growing recognition in Taiwan that the fight against AI-driven disinformation is not only a technological challenge but also a democratic one. It is a fight for the integrity of elections, for the quality of public debate, and for the willingness of citizens to engage with one another across political divides without falling into traps of hatred and misunderstanding.
Ultimately, the picture painted by these experts is both sobering and motivating. AI is not inherently evil; it is a tool. But in the hands of those who seek to destabilize democracies, it becomes a powerful weapon—capable of spreading falsehoods faster than truth, deepening divisions that already exist, and convincing citizens that their voices do not matter. Taiwan is not alone in facing this threat; democracies around the world are grappling with the same challenges, from the United States and Europe to Southeast Asia. But Taiwan’s unique geopolitical position makes it a testing ground for techniques that may later be deployed elsewhere. What happens in the island’s elections could serve as an early warning system for the rest of the democratic world. The response, therefore, must be equally global in scope—sharing intelligence, coordinating content moderation efforts, and building partnerships that transcend national borders. Yet at the heart of the fight are individual citizens armed with curiosity, discernment, and a willingness to defend the truth. The experts in Taipei offered no easy solutions; there are none. But they reminded us that democracy is not a machine that runs by itself. It requires daily vigilance, conscious choices, and a resilient public that refuses to be manipulated. The human element—our capacity for empathy, critical thought, and genuine connection—remains the most powerful countermeasure to any algorithm, no matter how sophisticated. In the end, the battle for Taiwan’s information space is not just about winning an election. It is about preserving the possibility of an open society where facts still matter, where disagreement is possible without dehumanization, and where people can make decisions based on reality rather than illusion. That, more than anything, is worth defending.