Chinese AI Models Spearhead Annexing of Taiwan: Experts
Udumbara
.
Chinese artificial intelligence (AI) models are actively driving Beijing’s “unrestricted warfare” in the form of a cognitive campaign designed to systematically dismantle Taiwan’s social stability and democracy, with the ultimate goal of imposing a new global order, analysts warn.
The NSB examined five AI tools—DeepSeek, Doubao, Yiyan, Tongyi, and Yuanbao—and urged the public to heighten their awareness to safeguard their personal data and sensitive business information.
From Data Risks to Mass Chaos
Lo Cheng-chung, a professor with the Graduate Institute of Science and Technology Law at the National Kaohsiung University of Science and Technology in Taiwan, noted that while such security risks are not new, the growing adoption of Chinese AI tools among Taiwanese users has enabled the Chinese Communist Party (CCP) to collect and analyze extensive personal information on an unprecedented scale.“When you register for expanded features on a Chinese AI platform, your personal data is gradually compiled through big data analytics, giving the CCP a clear view of your personal preferences, home address, and financial situation,” Lo told The Epoch Times.
Lo warned that, as the CCP represents Taiwan’s greatest national security threat, this complete exposure of citizen information creates a “comprehensive risk” for Taiwanese society that could give Beijing an advantage in a potential invasion.
“Imagine a flying robot dog landing, transforming into a humanoid, and hunting people with a machine gun—this is not science fiction but technologically feasible today, and modern warfare is deeply intertwined with technology,” Lo said.
“If Chinese AI platforms continue to gather intelligence on our populace, this data could enable such weapons to target citizens with precision during an invasion, which is extremely unsafe for Taiwan.”
“The true strategic value of these tools lies in generating mass-scale impact. Once the CCP harvests data from tens of thousands of people, they could theoretically link to connected devices—like overcharging the batteries of Xiaomi robot vacuums to ignite simultaneous fires across Taipei—creating immediate panic,” Lee told The Epoch Times.
“Furthermore, the algorithms can feed you fabricated content or even mimic Taiwan government notifications to manipulate perception, ultimately triggering catastrophic social subversion.”
.
Disinformation Orchestrates Peace Pact
In addition to significant data security risks, the NSB discovered that the five models’ generated content contained serious distortions and disinformation, including adopting the CCP’s official stance that “Taiwan is governed by China’s central government”—a claim that contradicts Taiwan’s status as a self-governing democracy.The generated content also excluded keywords such as “democracy” and “Tiananmen Square,” indicating that the material had undergone political censorship.
The NSB further noted that the China-based AI models can easily generate highly inflammatory content that defames individuals and spreads rumors, raising concerns that such capabilities could be used for propaganda purposes in information warfare.
Lee said that the spread of such disinformation reflects the CCP’s tactics of “unrestricted warfare,” where these AI models are used to ensure that Taiwanese users only see information that benefits Beijing, while also pushing its non-military efforts to gradually influence Taiwan and bring it under CCP control.
“On the surface, information obtained from Chinese AI appears objective, but in reality it all originates from a CCP-controlled system, creating an isolated universe within the Chinese-language sphere. This is the greatest danger,” Lee stated.
Lee added that once these Chinese AI models establish trust among local users, they are positioned to amplify “U.S. skepticism”—the propaganda narrative that the United States is an unreliable partner—to foster pro-China sentiment.
.
“They will flood Taiwan with messaging that characterizes normal Taiwan-U.S. trade as evidence that we are merely American pawns or puppets, aiming to erode the public’s psychological defense against China,” Lee explained.
Lee observed that Beijing could take this manipulation further by using AI to generate content that initially praises Taiwan to feign neutrality before subtly pivoting to pro-China narratives, ultimately aiming to seize the island without military force.
Pre-War Subversion
Echoing Lee’s concerns, Lo said that the CCP’s deliberate propagation of “U.S. skepticism” narratives threatens to erode Taiwan’s liberal democracy in favor of China’s socialist democracy.“These models weaponize their sophisticated language capabilities to entrap users, spreading defeatist narratives arguing that Taiwan should not waste money on arms purchases—a path that will inevitably turn Taiwan into the next Hong Kong,” Lo said.
Expanding on the threat, Lee cautioned that even if sowing defeatism fails to force an immediate surrender, Beijing would fully leverage these AI models to secure a comprehensive advantage over Taiwan before launching a kinetic attack.
“The AI can cripple the government’s authority by stoking public distrust, claiming that government communications are flawed, outdated, or concealing the truth,” Lee said.
Beyond undermining governance, Lee added that the CCP could utilize these tools to trigger resource shortages and economic instability.
“They could incite panic buying of essentials like toilet paper and cooking oil or spread disinformation about transportation gridlocks, aiming to destabilize civil society prior to an attack,” Lee said.
Lee predicted that once social order collapses, Beijing would continue using AI to rally support for pro-CCP legislators and political parties to pave the way for a total takeover of Taiwan.
“In a worst-case scenario, a compromised parliament could establish a provisional government or proxy regime, making Taiwan the first casualty in Beijing’s ambition to impose its own global order,” he said.
.
