When Shadows Speak: Taiwan’s Battle Against AI-Powered Disinformation

On a tense morning in Taipei, a gathering of scholars, political strategists, and concerned citizens took their seats inside the Democratic Progressive Party headquarters, acutely aware that the clock was ticking toward election day. The air was thick with questions that had no easy answers: How do you defend a democracy against an adversary that no longer needs to shout to be heard? How do you fight back against weapons that are invisible, automated, and endlessly adaptable? This was the heart of the conversation at the DPP’s forum on foreign electoral interference, where experts had been summoned to dissect a threat that feels both futuristic and disturbingly immediate. The message that emerged was stark but clear: the battlefield of the next election will not be won with tanks or missiles, but with pixels, posts, and perfectly crafted half-truths designed to slip past our defenses before we even know they are there.

The warning that set the tone for the entire forum came from DPP Secretary-General Hsu Kuo-yung, who articulated what many in the room already suspected but had struggled to put into words. China, he explained, is not interested in outright lies that can be easily debunked. Instead, Beijing has perfected a formula that is far more insidious: information that is thirty percent true and seventy percent false. This hybrid approach is brilliantly deceptive, because the kernel of truth gives the falsehood a sheen of credibility, making it nearly impossible for the average person to distinguish fact from fabrication. Professor Yeh Yao-yuan, chair of international studies at the University of St. Thomas in Houston, expanded on this idea, offering a sobering analysis of what Beijing actually hopes to achieve. In the short term, the goal is simply to influence the outcome of specific elections. But the long-term objective is far more corrosive: to deepen social divisions, to make citizens distrust their institutions, and to slowly erode the very foundation of democratic trust. This is not about winning a single contest, Yeh cautioned, but about making democracy itself seem broken, chaotic, and unworthy of belief.

Perhaps the most chilling moment of the forum came when Yeh shared a piece of intelligence that had recently emerged from Washington. In July, the United States government discovered that China had successfully stolen over 220 million American voter registration records. That is not a hypothetical scenario or a theoretical vulnerability; it is a demonstrated capability. With that vast repository of personal data, Yeh explained, China’s analysts can identify precisely which issues resonate with voters in different regions, allowing them to craft targeted disinformation campaigns that exploit existing anxieties and grievances. The implication for Taiwan was unmistakable. If a nation with the cybersecurity resources of the United States could be breached in this manner, what chance does a smaller democracy have? The answer, Yeh suggested, is that Taiwan must urgently enhance its safeguards before it becomes the next victim of this data-driven warfare. The warning resonated deeply with the audience, many of whom suddenly understood that the fight against disinformation is not merely about fact-checking individual posts, but about protecting the most intimate details of citizens’ lives from being weaponized against them.

The conversation then turned to the specific tools Beijing is now deploying, and here Professor Liu Yu-hsi from Shih Hsin University offered a glimpse into the strange new world of AI-generated interference. The shift toward artificial intelligence, she explained, has fundamentally changed the economics of disinformation. Where once it required teams of writers, translators, and graphic designers to produce propaganda, now a single operator can generate massive volumes of text and images in a matter of minutes. To illustrate this point, Liu pointed to a curious phenomenon that had emerged during the recent local elections: several Chinese immigrants had registered to run for borough warden positions, a low-level but symbolically significant office. What made these candidacies suspicious was not the individuals themselves, but the campaign materials they produced. Their online graphics were riddled with errors, strange grammatical constructions, and visual inconsistencies that suggested they had been generated by artificial intelligence rather than crafted by human hands. These were not serious campaigns, Liu argued, but experiments in testing the boundaries of what AI can accomplish in the political sphere. The sheer volume of such content, she warned, threatens to overwhelm our ability to process information critically, creating a fog of confusion in which voters can no longer discern what is real and what is manufactured.

Arthur Wang, secretary-general of the Asia-Pacific Elite Interchange Association, provided the historical context that helped explain how China arrived at this point. His account traced the evolution of Beijing’s tactics over several decades, revealing a strategic mind that is constantly adapting and learning. In the early years, China relied on blunter instruments: verbal attacks from official spokespeople and periodic military threats designed to intimidate. By 2008, the approach shifted toward economic inducements, using trade and investment to create dependencies that could be leveraged for political gain. Then, between 2016 and 2018, the focus moved decisively toward disinformation, as Beijing discovered the power of social media to shape public perception. Now, Wang argued, China has entered a more sophisticated phase characterized by targeted election interference. This is not indiscriminate propaganda but surgical precision, aimed at identifying and amplifying existing fault lines within Taiwanese society. The goal is to make the divisions between different groups of citizens so deep and so bitter that they become the dominant feature of the political landscape. When a society is busy fighting itself, Wang observed, it has little energy left to defend its sovereignty. This understanding of China’s long game is crucial, because it reveals that the threat is not just to individual elections, but to the very fabric of social cohesion.

The final portion of the forum shifted from diagnosis to prescription, as Wu Jun-zhi, director of the DPP’s China Affairs Department, offered practical guidance for ordinary citizens navigating this treacherous information landscape. His advice was refreshingly concrete, a useful antidote to the overwhelming scale of the problem. Be suspicious, he urged, of information that comes from unknown sources or that seems designed to provoke an emotional response. Disinformation thrives on outrage and fear, because these emotions short-circuit our critical thinking faculties. When you feel your blood pressure rising at a particular post or article, that is exactly the moment to pause and verify. Wu recommended a simple protocol: check the source of the information, cross-reference it with reputable media outlets, and take a moment to consider whether the content is trying to make you feel something rather than inform you. On a broader level, he joined his fellow panelists in calling on the government to invest in AI databases and filtering systems that can help identify and flag suspicious content before it spreads too far. But perhaps the most important recommendation was the emphasis on media literacy education. Citizens must be taught not just how to consume information, but how to question it, how to trace its origins, and how to recognize the patterns of manipulation. In the end, the forum concluded, the most powerful defense against disinformation is not technology but an informed and skeptical public. In the battle for the future of democracy, the greatest weapon we possess is our own willingness to think critically, to demand evidence, and to refuse to be divided by those who profit from our confusion.