China has spent the past decade intensifying pressure on Taiwan through a range of actions that deliberately stop short of a traditional military invasion. These so-called gray zone tactics encompass diplomatic, economic, and security measures, alongside the regular deployment of ships and aircraft into areas that Taiwan considers its maritime zones and airspace. Individually, each maneuver may appear minor, but their collective effect over time piles immense pressure on the island-nation, threatening its democratic foundations without requiring a single shot to be fired. Central to this entire strategy is China’s focus on cognitive warfare, which aims to manipulate the perceptions and decision-making processes of the Taiwanese public, ultimately eroding the island’s will to resist and fostering conditions favorable for reunification.
The concept of cognitive warfare has its roots in China’s military modernization during the late 1990s and early 2000s, when the People’s Liberation Army developed the principles of “unrestricted warfare.” This doctrine proposes using all elements of national power, including economic, diplomatic, and informational tools, to achieve strategic objectives, moving beyond conventional battlefield confrontations. Western scholars simultaneously developed frameworks for power, including hard power, or military might and coercion, and soft power, which uses attraction and persuasion to influence a target population. Sharp power, a related concept, employs similar methods to soft power but with the intent to confuse and demoralize rather than persuade. Within the context of unrestricted warfare, information warfare emerges: the deliberate use of information networks, like the internet, to gain a competitive advantage through propaganda or social media bots. The most direct and potentially disruptive form of information warfare, however, is cognitive warfare, a term coined around 2017 to describe the act of manipulating an enemy’s cognition. Unlike other information-related strategies that aim to make an enemy believe a certain fact, cognitive warfare seeks to impede decision-making, polarize society, weaponize identity, and erode the will to fight by creating distrust and confusion.
The specific case of Taiwan illustrates these tactics in action. For instance, Chinese-controlled bots on Taiwanese social media platforms are not primarily focused on promoting a singular pro-China narrative. Instead, their goal is to sow distrust, drive a wedge between the Taiwanese people and their government, and encourage political paralysis. This approach is most visibly used during election cycles, where Chinese-linked actors actively target the independence-leaning Democratic Progressive Party with rumors and conspiracies, aiming to delegitimize its leadership and undermine public faith. According to organizations like Doublethink Lab, a Taiwanese civil society group, China’s 2020 election interference included propaganda and content farms orchestrated by entities like China’s Propaganda Department, People’s Liberation Army, and United Front Work Department. These state bodies set the messaging, but the actual dissemination is often carried out via the Chinese People’s Political Consultative Conference and the Taiwan Affairs Office, which directs content farms to spread disinformation. This decentralized network has remained consistent, but the tools and technologies it employs have evolved, especially with the rise of artificial intelligence, leading to a shift from text-based posts to AI-generated or manipulated videos.
China’s strategy has also adapted to the changing social media landscape, moving away from content farms and toward amplifying pre-existing stories rather than inventing new ones from scratch. For example, the claim that the US is hollowing out Taiwan Semiconductor Manufacturing Company, or will abandon Taiwan once production shifts to the United States, was a narrative that originated in Taiwan, but UFWD-linked media groups quickly and strategically pushed to exaggerate these fears, aiming to sow discord. Chinese entities are not trying to create new false stories but rather exploit and magnify already existing anxieties within Taiwan’s open society. This tactic is also key to influencing public opinion through non-social media avenues. DPP legislator Puma Shen’s testimony highlighted three approaches: direct, through CCYS propaganda; indirect, via financian back bets. Taiwan’s social media is bombarded by AI-driven content that amplifies general security fears or mimics local political narratives, according to analysis from Doublethink Lab. They also found a correlation between the information sources Taiwanese citizens use and their political beliefs, with supporters of the opposition parties more likely to agree with narratives promoted by Chinese-linked accounts.
Looking at the future, China is already preparing for Taiwan’s local elections in 2026 and the presidential election in 2028, tailoring its approach to the increasingly fragmented and AI-driven information environment. The leaked documents from the Chinese company GoLaxy reveal a precision data-scraping system and “Smart Propaganda” technology that leverages publicly available social media data to create tailored content in real-time. GoLaxy tracks specific internet users and political situations to push targeted messages. In December 2023, moments before Taiwan’s 2024 election, the company was recommending specific messaging to exploit public divisions, claiming to have produced political effects in state departments that could bolster China’s voice. This shows a significant shift in strategy. Chinese information is no longer just about broadcasting a single message; it is about building a comprehensive AI-powered ecosystem that grants individuals with tailored disinformation. As local elections are particularly vulnerable to fictional fake news, China has already started planning for 2026, with directives to use its united fund to weaken Taiwan independence forces, and attempts to achieve the “weaponize” Taiwanese identity through the election process are already underway.
In the lead-up to the 2026 and 2028 periods, this offensive is likely to accelerate. Taiwan’s National Security Bureau has already detected multiple suspicious accounts and disputed messages tied to AI–generated content on foreign affairs, national defense, and the economy. Another wave came in the wake of the Trump Xi summit, with nine thousand messages pushing the narrative that Taiwan can’t trust the US and reunification is the only path to peace. This campaign is a trying test on two democracies in the world. Chinese campaigns have evolved a nasty web where the goal is not to persuade but to stoke divisions, erode trust, provoke public frustration, and destabilize the social fabric of Taiwan. While Taiwan’s open society and polarizing political environment make it vulnerable, its high tech, smart population, and a public that is increasingly aware of disinformation campaigns provides a level of resilience. This resilience, combined with broad-technological advantages and collaborations with partners like the US, can create more resilient democratic systems. The outcome will determine whether AI is used to erode the world’s democratic order or strengthen the determination and security of a free and open society.
