For the past decade, Beijing has been waging a quiet, relentless siege against Taiwan—one that deliberately avoids the thunder of artillery and the roar of fighter jets. Instead, this pressure manifests through a dizzying array of gray zone tactics: diplomatic isolation, economic coercion, and the constant, unnerving presence of Chinese ships and aircraft probing the edges of Taiwan’s maritime and airspace boundaries. Each individual action might seem minor, a mere bureaucratic or territorial pinprick, but their cumulative effect is a slow, grinding erosion of the island’s democratic foundations. Central to this campaign is a sophisticated weapon that doesn’t require a single shot to be fired: cognitive warfare. This isn’t just about spreading lies; it’s about fundamentally altering the way a population thinks, trusts, and decides. By targeting the very cognitive processes of Taiwanese society, Beijing aims to weaken the island’s will to resist, polarize its politics, and ultimately pave the way for reunification on China’s terms, all while maintaining the plausible deniability of a nation technically at peace. The strategy is insidious precisely because it operates in the shadows of normal political discourse, making it difficult to attribute, counter, or even fully comprehend until the damage is already done.
To understand this modern threat, one must trace its intellectual lineage. In the 1990s and early 2000s, China underwent a profound “revolution in military affairs,” modernizing the People’s Liberation Army and, more importantly, its strategic thinking. This era birthed the concept of “unrestricted warfare,” a doctrine that embraces all domains of conflict—economic, diplomatic, informational, and military—to achieve strategic advantage without necessarily resorting to traditional combat. Simultaneously, Western scholars were refining their own frameworks of power. Hard power relies on military might and coercion; soft power uses attraction and persuasion through culture and diplomacy. But there is a darker variant: sharp power. While it mimics the mechanisms of soft power, its intent is not to persuade but to confuse, demoralize, and manipulate. Within this nexus of unrestricted warfare and sharp power lies information warfare—the deliberate use of the internet and media to gain a competitive edge. This can range from state-run propaganda to swarms of social media bots. However, the most modern and insidious iteration is cognitive warfare. Formally articulated around 2017, cognitive warfare distinguishes itself by its ultimate objective. It doesn’t merely seek to make an enemy believe a specific falsehood. Instead, it weaponizes disinformation to impede decision-making, polarize society, weaponize identity, and erode the very will to fight. For instance, Chinese-controlled bots on Taiwanese social media aren’t primarily trying to sell a pro-China narrative; they are sowing distrust between the Taiwanese people and their government, driving a wedge between citizens and their democratic institutions. This strategy flares up most intensely during election cycles, targeting the independence-leaning Democratic Progressive Party (DPP) with a relentless barrage of rumors and conspiracies designed to delegitimize their governance and cast doubt on the integrity of the democratic process itself.
The empirical evidence of this campaign is meticulously documented by Doublethink Lab, a Taiwanese civil society organization dedicated to studying disinformation. Their 2021 report, “Deafening Whispers,” dissected China’s efforts to influence the 2020 presidential election, revealing a complex, decentralized network. The Chinese Communist Party (CCP) sets the overarching narrative, but the actual dirty work is outsourced. The Propaganda Department, People’s Liberation Army, and United Front Work Department (UFWD) dictate the themes, while entities like the Chinese People’s Political Consultative Conference and the Taiwan Affairs Office direct a sprawling network of “content farms” to disseminate the material. Over the past six years, the cast of characters has remained remarkably stable, but the technology has evolved dramatically. By the 2022 local elections and the 2024 presidential race, Doublethink Lab observed a significant shift. As social media platforms cracked down on coordinated inauthentic behavior, China relied less on low-quality content farms and pivoted to a more sophisticated approach. The proliferation of artificial intelligence (AI) meant that disinformation transitioned from text-based posts to AI-generated or manipulated videos. More importantly, the strategy shifted from creating false stories to amplifying existing ones. A prime example is the rumor that began circulating in December 2021 about the United States hollowing out Taiwan’s crown jewel, Taiwan Semiconductor Manufacturing Company (TSMC), and abandoning Taiwan’s defense once production moved stateside. While this narrative originated organically in Taiwan, UFWD-linked media groups quickly seized upon it, exaggerating claims that the DPP was selling off the semiconductor industry for its own benefit. This is the hallmark of modern cognitive warfare: not inventing a lie, but taking advantage of an open society’s existing fears and amplifying them to a fever pitch, stoking partisan divisions and eroding trust in the government.
The mechanics of this manipulation are deeply intertwined with Taiwan’s political landscape. Doublethink Lab’s phone surveys have established a strong correlation between information consumption and belief in Chinese-linked narratives. China’s information manipulation specifically targets supporters of the opposition Kuomintang (KMT) and the Taiwan People’s Party (TPP), aiming to spread anti-DPP messaging. While partisan bickering is a staple of any healthy democracy, Chinese bots are distinguishable by their synchronized posting times, view-botting, and the use of phrases unique to mainland China. These accounts infiltrate online KMT and TPP echo chambers, spreading disinformation about Taiwan, DPP leadership, and cross-strait relations. The narratives first gain traction within these closed communities before leaking out to the broader public. Survey data shows that KMT and TPP supporters are significantly more likely to agree with these Chinese-linked rumors, while DPP supporters are not. Furthermore, individuals who rely on social media influencers, online content creators, and fan pages—all heavily targeted by Chinese campaigns—tend to align more closely with China’s narratives. While causation is difficult to prove, the correlation is stark. The insidious genius of this approach is that the end goal isn’t to make the audience believe a specific story, but to distort their overall perception of the DPP and corrode their faith in the government. Each individual conspiracy might be dismissed, but their cumulative effect is to undermine the legitimacy of political debate itself. This is further compounded by other vectors of influence. DPP legislator Puma Shen testified that China uses media, academia, technology, diplomacy, and the military to interfere in information environments. He outlined three strategies: direct manipulation (CCP propaganda and social media), indirect investment (financing marketing companies and influencers, leading to self-censorship in Taiwanese media), and ideology-driven recruitment (bringing Taiwanese citizens critical of their government into private chat groups to spread manipulated content). It’s worth noting that Shen and Doublethink Lab have ties to the DPP, which speaks to the partisan nature of the fight against disinformation, but it does not invalidate their findings.
Looking toward the future, the threat is escalating exponentially. In August 2025, Vanderbilt University obtained leaked documents from the Chinese company GoLaxy, which scrapes social media data to analyze public sentiment for corporations and the Chinese government. The documents revealed GoLaxy’s “Smart Propaganda System” (GoPro), a technology that collects detailed information on specific internet users to produce targeted content in real time. In December 2023, just a month before Taiwan’s presidential election, GoLaxy made specific recommendations on messages to push to exploit divisions in Taiwanese public opinion. The company boasts that it has “already produced certain political effects in relevant state departments,” with the ultimate goal of making “the easterly wind overpowering the westerly wind” a reality. This represents a seismic shift from broadcasting a single message to micro-targeting individuals and small groups. AI has made this precision possible, moving the focus away from fact-checking individual rumors to countering an entire AI-powered ecosystem of tailored disinformation. The plans for Taiwan’s 2026 local elections are already in motion. Top CCP officials met in February during the Taiwan Work Conference to establish a task force specifically to interfere in the upcoming elections, using the united front to weaken “Taiwan independence forces.” Local elections are particularly vulnerable due to their fragmented information space. Already, Taiwan’s National Security Bureau (NSB) has detected two waves of attacks in 2026: one in April involving 13,000 suspicious accounts and 860,000 disputed messages containing AI-generated content on foreign affairs and defense, and another following the Trump-Xi summit in May, pushing the long-standing “American Skepticism” narrative—the idea that Taiwan cannot trust the United States and that only reunification will bring peace.
As AI technology continues to improve, the challenge for Taiwan becomes increasingly daunting. The government, civil society, and analysts are in a constant arms race against a sophisticated, evolving adversary. Chinese cognitive campaigns have moved beyond state media narratives into a sprawling web of disinformation and amplification designed to stoke internal divisions. Yet, Taiwan is not a passive victim. It is a highly literate society, and years of persistent attacks have heightened public awareness of disinformation tactics. Taiwan’s formidable advantages in AI expertise and technology production can be harnessed, particularly in collaboration with democratic allies like the United States, to build more resilient information ecosystems. The coming years will serve as a global case study. Will AI technology fundamentally undermine democratic governance, or can open societies adapt and safeguard themselves? The answer hinges on whether Taiwan can leverage its technological prowess and international alliances to protect its information environment. China’s cognitive warfare against Taiwan is not just a regional issue; it is a test run for democracies worldwide, highlighting both the terrifying potential of AI in new domains of conflict and the urgent need for a coordinated, democratic response to preserve the integrity of public discourse and the very foundation of self-governance. The battle for Taiwan’s future is being fought not on a physical battlefield, but in the minds of its citizens—and the outcome will resonate far beyond the island’s shores.