Cognitive Warfare: The AI Industrialization Phase of Taiwan's Information Battlefield

30-Second Overview:
CCP cognitive warfare against Taiwan entered a new phase of "AI Industrialization × Using Taiwan to Criticize Taiwan × Named Retaliation" between 2024 and 2026. In December 2024, Ba Jiong and Chen Po-yuan released the first part of the China United Front Documentary, exposing Fujian Province's Taiwan-focused united front bases, with the first part surpassing 2 million views¹. One year later, in December 2025, an AI-generated fake doctor named "Chen Chih-ming" impersonating staff from National Taiwan University (NTU) and Taipei Veterans General Hospital (TVGH) appeared, prompting TVGH Vice President Li Wei-chiang to personally clarify the situation². A Reuters and IORG investigation found that in Q4 2025, 1,076 CCP state media accounts published 560,000 videos on Douyin, using facial recognition to identify 57 Taiwanese individuals appearing in the content³. However, this is a two-way battlefield: in October 2025, the Chongqing Public Security Bureau filed a case against Legislator Shen Pei-yang for the "crime of splitting the country" under the charge of "splitting the nation"⁴. At the height of this counter-narrative, UNLV Political Science Assistant Professor Wang Hong-en (Austin Wang) warned: "A group of accounts that always log on together for warmth might just really be a group of good friends"⁵. This entry attempts to map a three-sided cognitive battlefield that requires vigilance against CCP manipulation, vigilance against local labeling, and an acknowledgment of the limitations of research tools themselves.

An Elderly Person’s LINE Group

On December 28, 2025, Li Wei-chiang, Vice President of Taipei Veterans General Hospital (TVGH), received inquiries from the media: Who was the man in the white coat on the YouTube channel "Dr. Chen Chih-ming," speaking with certainty and claiming to be a doctor at TVGH?²

The answer was: No such person exists.

Li Wei-chiang told reporters that "after verification, the hospital confirmed there is no such doctor." The channel even featured two completely different faces for "Chen Chih-ming"². The perpetrator used an AI-generated virtual avatar to impersonate a medical center doctor from TVGH and NTU, attracting elderly users to share videos with pseudoscientific claims such as "treating diabetes with bacterial infections." The Taiwan FactCheck Center (TFC), after technical analysis, pointed out that this was an AI video generated by a "multimodal, audio-driven character animation generation model"². By the time it was exposed, the channel had accumulated 23,100 subscribers and a single video had reached 150,000 views⁶.

📝 Curator’s Note: The most critical detail in this case is that it could be named and clarified: The Vice President spoke directly, TFC published a public fact-check, and the Ministry of Health and Welfare announced regulatory oversight². This entire structure of accountable response is something the Chinese side of Douyin cannot replicate. Verifiability itself is the texture of a democratic society.

In the same year, the New Taipei City Social Affairs Bureau received calls from citizens asking how to apply for "monthly subsidies of over 8,000 NTD for the elderly." Verification revealed it was another AI-generated video circulating on YouTube and LINE groups: the character’s expressions were unnatural, pauses in speech were illogical, and the tone lacked genuine emotion⁷. Artist Wang Jen-fu’s family experienced voice deepfake fraud: his daughter first received a fake market research call to collect voiceprints, and days later, his wife Chi Chin received a distress call from "the daughter" demanding a wire transfer⁸. The Criminal Investigation Bureau later broke the first AI voice fraud case in Taiwan, revealing the industrialized nature of the operation: an illegal investment advisory group designed over 20 dialogue scenarios, deceiving at least 70 people, with one woman losing 20 million NTD⁹.

The common thread in these three scenarios points to the same thing: making you "stop believing."

Scale: From Cases to Systems

The National Security Bureau’s 2025 report, Analysis of CCP Cognitive Warfare Tactics Against Taiwan, revealed that in that year, over 45,000 abnormal accounts were reported, over 2.31 million pieces of contentious information were collected, and over 3,200 reports were made to government agencies¹⁰.

A Reuters and Taiwan Information Environment Research Center (IORG) investigation published in April 2026 showed that in Q4 2025, 1,076 accounts operated by CCP state media published approximately 560,000 videos on Douyin, of which approximately 18,000 involved Taiwan-related topics³. By cross-referencing 2,730 such videos with facial recognition, 57 Taiwanese individuals were identified as appearing in the content; among the top 25, 13 were associated with the Kuomintang (KMT). The most exposed figure was then-KMT Chairperson Cheng Li-wen: she appeared in 460 videos published on 68 Douyin accounts, accumulating over 5 million interactions³.

Another IORG tracking study indicated that among 84 "doubt-the-US" narratives, 70 were initiated or amplified by CCP state media¹¹. A 2025 study by Doublethink Lab (Taiwan Democracy Lab) indicated that heavy TikTok users were more likely to be pro-Beijing and more likely to认同 "populist skepticism" positions¹².

But volume itself does not equal "battlefield success." The observation of criminologist and co-founder of the Black Bear Academy, Shen Pei-yang, is worth including: "80% of the propaganda war China conducts against Taiwan has nothing to do with truth or falsehood; it is mostly narrative attacks, i.e., fabricating a perspective."¹³ This statement shifts the focus from the first reflex of "fact-checking" to a more difficult second layer: when the adversary is deploying perspectives and narrative frameworks, the reach of fact-checking is simply not far enough.

Looking back two election cycles further clarifies the axis. IORG tracking found that narratives like "ballot stuffing" first appeared in late 2019, reappearing repeatedly during the 2020 general election, the June 2020 recall of Han Kuo-yu, and the 2021 referendum period. The 2020 general election also saw a new phenomenon of direct intervention by Chinese officials, with attack targets even including the legitimacy of the Central Election Commission¹⁴. After Lai Ching-te’s election in 2024, the National Security Bureau report noted new characteristics: AI virtual anchors spreading disinformation about Tsai Ing-wen, deepfake videos, and election betting markets¹⁵. In 2025, AI-fabricated videos of Lai Ching-te "cursing" even appeared¹⁶.

Five Coordinated Tactics Dissected by the National Security Bureau

The National Security Bureau has broken down CCP cognitive warfare against Taiwan into five tactics[^10]:

The significance of this breakdown is replacing the vague term "foreign forces" with a structure where responsibility can be traced to the corporate level: each tactic corresponds to specific companies and government units within China. But this also means the complexity of fact-checking and prosecution has risen significantly: videos are mass-produced by offshore AI factories, reposted by anonymous local Taiwanese accounts, and then pushed back to Taiwanese readers by algorithms; the entire chain has no single speaker.

How Users "Read" a Suspicious Channel

In February 2025, Threads user @derek_foxx posted a thread¹⁷ documenting how he determined that the YouTube channel "Unicorn So Loud" (Dujiaoshu Name Jiao) was problematic:

1. Accent Clues: The speaker’s voice lacked a Taiwanese accent, deliberately catering to Taiwanese usage but revealing details in subtleties.
2. Transformation Traces: The channel’s early content focused on financial management and lifestyle habits (typical content farm material), shifting to "MythBusters"-style content about a year ago.
3. Wind Direction Synchronization: The content themes were identical in tone to livelihood criticisms from certain opposition camps.
4. Anonymity: The operator did not disclose their identity, later switching entirely to AI dubbing.

This warning was retweeted and discussed by TFC and other users¹⁸, pointing out that the channel had over 340,000 subscribers, focusing on livelihood-sensitive topics like "Taiwanese fresh milk is more expensive than Europe and the US," "Taiwanese tilapia is a renaming scam for tilapia," and "exchange rate manipulation causes low wages," using editing and decontextualization to amplify emotion¹⁹.

📝 Curator’s Note: This section requires careful handling. Taiwan.md is not an intelligence agency; we cannot arbitrarily accuse a specific channel of being directly operated by the CCP: as of the completion of this entry, no research institution or government unit has listed this channel as a certified foreign-operated case. The significance of retaining this section is to demonstrate how general readers can ask questions based on observable characteristics: accent, anonymity, algorithmic traces, and thematic shifts; these four clues combined can at least give people a pause button to "not rush to share." Determining who is a CCP account is the work of judicial and research institutions. Training one’s own literacy intuition is something everyone can do.

The AI Cost Cliff

How low has the tool threshold for contemporary cognitive warfare become? Before the 2024 New Hampshire primaries in the US, political operator Steve Kramer hired magician Paul Carpenter to synthesize former President Biden’s voice using ElevenLabs, making thousands of robocalls urging Democrats not to vote²⁰. The FCC later fined $6 million, and Kramer was charged with 13 counts of election suppression (felonies) and 13 counts of impersonating a candidate²⁰. The production cost? $1, 20 minutes, and Kramer paid Carpenter $150.

In January of the same year, AI pornographic deepfakes of Taylor Swift appeared on X, with a single post exceeding 47 million views, causing X to temporarily suspend search for the "Taylor Swift" keyword²¹. The incident pushed the US Congress to pass the TAKE IT DOWN Act in 2025, mandating the removal of non-consensual intimate images²², and in January 2026, the DEFIANCE Act passed the Senate unanimously, granting civil rights to deepfake pornography victims²³.

This cost curve means: cognitive warfare on future election battlefields can be launched without Chinese support: as long as someone is willing to spend $1. The focus of defense shifts from "specific sources" to the fact that "a certain class of tools has become ubiquitous."

Two Lines of Counterattack: The Lens and the System

Taiwan is not just passively taking hits. Since 2024, at least two distinct counterattack lines have emerged: one is influencer-style lens counterattacks, the other is the construction of academic and civil defense systems. Both have been met with precise retaliation from the CCP.

Lens Counterattack: Ba Jiong and Chen Po-yuan’s _China United Front Documentary_

On December 6, 2024, YouTubers "Shetu Diary Fun TV" Ba Jiong (Wen Tzu-yu) and former "Little Pink" Chen Po-yuan (Minnan Wolf PYC) released the first part of the China United Front Documentary. Chen Po-yuan went undercover in Fujian Province, documenting how the CCP’s "Taiwan Youth Entrepreneurship Bases" operate: interest-free startup funds, assistance in obtaining "Chinese ID cards," and bank over-lending¹. The first part accumulated over 2 million views. The second part, released on the 28th of the same month, further exposed details of the Xiamen entrepreneurship base, surpassing 1.17 million views²⁴. Ba Jiong pointed out in the film that the CCP absorbs Taiwanese people who obtain Chinese ID cards through such programs, claiming the number has reached 200,000 (this is a claim in the film; the Mainland Affairs Council only stated it would "strictly investigate united front tactics," without directly confirming the specific number)²⁵.

The cost of the narrative arrived immediately. Chen Po-yuan’s Weibo was attacked by "Little Pinks," who called him a "Taiwan independence dog." His Douyin account "Minnan Wolf PYC" was deleted by Chinese authorities. He had participated in producing The Old Boss, which accumulated 1 billion views in the China region, but he never received copyright fees²⁴. The tension in this story is: anti-CCP figures were once products of the CCP; neither side of the strait easily welcomes "former Little Pinks."

System Counterattack: Shen Pei-yang’s "Enemy Medal"

The other line is the construction of academic and civil defense systems. The Black Bear Academy was founded in 2021 by Shen Pei-yang and Ho Cheng-hui, with honorary chairman of United Microelectronics Corp. (UMC), Tsao Hsing-cheng, donating 600 million NTD in 2022²⁶. IORG was co-hosted by Shen Pei-yang and Yu Chih-hao in 2019, publishing the annual Taiwan Information Environment Report to track the spread of narratives like "doubt-the-US"²⁷.

Retaliation followed the same axis. On October 14, 2024, the Taiwan Affairs Office listed Shen Pei-yang on the list of "stubborn independence separatists," banning him and his family from entering China, Hong Kong, and Macau²⁸. On October 28, 2025, the Chongqing Public Security Bureau issued a police bulletin, formally investigating Shen Pei-yang for the "crime of splitting the country," alleging he engaged in criminal activities splitting the country through the "Black Bear Academy," based on the Criminal Law of the People's Republic of China⁴. CCTV subsequently released an nearly 8-minute "exposé" special, threatening global arrest through Interpol. Shen Pei-yang responded: "One country, one side; don't妄想 (dream) of extending your hand into Taiwan."²⁹ (Note to readers: CCTV’s threat to arrest is program commentary and the statement of Professor Cheng Lei from the Law School of Renmin University of China, not an official arrest warrant from the Chinese Ministry of Public Security or the courts.)

This is the first time in Taiwan’s history that a legislator has been formally investigated by the CCP with a named charge⁴. Columnist Huang Peng-hsiao titled the piece: "Congratulations to DPP Legislator Shen Pei-yang on Receiving the 'Enemy Medal.'" The irony is that the person teaching Taiwanese how to conduct cognitive warfare was certified by the CCP’s state machinery as having value in his work.

The contrast between these two lines reveals differences in counterattack strategies (Ba Jiong’s lens relies on emotional penetration and personal代言, while Shen Pei-yang’s system relies on academic research and civil defense training), as well as the precision of CCP retaliation: deleting accounts for influencers, filing cases for system builders.

Beyond the CCP: Local Shadows, Platform Responsibility, and Lexical Tension

Beyond defense, an honest observation must acknowledge: Taiwan also has shadows of information operations internally, and tensions regarding the potential abuse of the term "cognitive warfare."

Local Cyber Armies: Lin Wei-feng, 1450, and the "47 Accounts" Blunder

In 2021, pro-DPP writer Lin Wei-feng was exposed for using the account bj26bj on PTT to post pro-CCP articles in reverse, then using his real name on Facebook to accuse PTT of being penetrated³⁰. The Investigation Bureau received over 70 reports, and the incident burned for months. Lin Wei-feng’s spouse, Yang Min-shi, was then the Deputy Director of the DPP’s Network Community Center³⁰.

The term "1450" refers to the DPP’s cyber army sidekick. The official version is that the Council of Agriculture allocated 14.5 million NTD for network marketing in 2019, hence the name. The opposition version directly points to the DPP’s organized cyber army operations³¹. Another case: the DPP administration once listed 47 network accounts as Chinese "agents," only to find after verification that most were DPP supporters³².

The Kansai Airport Incident: A Historical Textbook for the Truth Triangle

The 2018 Kansai Airport incident is the most difficult but most reference-worthy case of "cognitive warfare boundaries" to date.

On September 6 that year, Typhoon Jebi hit Kansai Airport. Subsequently, false information circulated on the Taiwanese internet claiming "the Chinese Embassy sent cars to pick up Chinese passengers," criticizing the Taiwan Representative Office in Osaka for inaction. Shen Pei-yang later traced the source, confirming the fake video first appeared on the Chinese Weibo account "Floods and Fierce Beasts baby," made into news by "Observer Network," 24 hours earlier than Taiwanese news³³.

But this was only half the story. On the Taiwanese local side, Yang Hui-ju provided internet to Cai Fu-ming, who used the PTT account idcc to criticize the Osaka Representative Office as "rotten to the core, remnants of the party-state state." Within one minute, he directed the cyber army via the LINE group "Kaohsiung Group" to amplify the tone³⁴. On September 14, the Director of the Taipei Representative Office in Osaka, Su Chi-cheng, committed suicide at his official residence. His suicide note did not directly mention pressure from fake news, only writing "did not want to be humiliated" and "unwilling to be punished, transferred, or recorded for demerits due to baseless charges"³⁵. In November 2021, Yang Hui-ju was sentenced to 6 months in the first instance³⁶. She later applied for constitutional interpretation; in March 2025, the Constitutional Court ruled the crime of insulting official duties unconstitutional, leading to Yang Hui-ju’s case being dismissed³⁷.

📝 Curator’s Note: The value of this case lies in laying bare the boundary between "cognitive warfare" and "local cyber armies." The same event can have both Chinese-source disinformation and relayed Taiwanese local cyber armies; any simplification is disrespectful to the deceased. Su Chi-cheng’s family has never been able to recover, and the Constitutional Court’s judgment made the legal responsibility dismissal final: justice, truth, and pain, these three lines cannot align.

Austin Wang’s Warning: Three Elements of Cognitive Warfare

In a long essay on Medium, UNLV Political Science Assistant Professor Austin Wang (Wang Hong-en) explicitly lists the three elements of "cognitive warfare": coming from abroad, coordinated attacks, and having specific motives: all three are indispensable⁵. He also warns:

"A group of accounts that always log on together for warmth might just really be a group of good friends."

"Messages transmitted by people of different colors are judged to be disinformation with different probabilities."

Wang’s stance is not to exonerate the CCP: he himself studies CCP cognitive warfare. He points out the risk of methodological double standards: academic research mostly only has access to public data, unable to access financial flows and communications. The government, acting simultaneously as the punisher and the information holder, has an agency problem. People First Party Legislator Chang Chi-lu also said something similar: "Treating all critics of the government as CCP accomplices or directed by the old CCP will only make people feel this is a green camp cognitive warfare operation."³⁸

Why the Digital Intermediary Services Act Was Withdrawn

On June 29, 2022, the National Communications Commission (NCC) passed the draft Digital Intermediary Services Act. At the third public hearing on August 18, Meta, Google, and PTT strongly reacted, arguing the act stifled freedom of speech. On August 20, Premier Su Tseng-chang intervened to postpone the hearing. On September 7, the NCC announced the withdrawal of the draft for internal review, with no timeline³⁹. The core concern of the opposition was: the platform responsibility mechanism might cause operators to delete content excessively to avoid liability, becoming "speech gatekeepers."

This withdrawal is the most concrete marker of the tension between "cognitive warfare governance" and "freedom of speech." As of April 2026, the Digital Intermediary Act has not been resubmitted.

Meta’s 16 Billion Threshold

Facebook’s platform responsibility is also worth looking at. A Reuters exclusive investigation in November 2025 revealed that Meta internally estimated approximately $16 billion (10% of global revenue) in 2024 came from scam ads⁴⁰. Daily push of "high-risk" scam ads reached 1.5 billion, with annual revenue of $700 million. Meta only bans when the automated system is 95% certain it is a scam, because widespread introduction of advertiser verification would reduce revenue.

But there is a detail that is a positive signal for Taiwan: Meta only introduced advertiser verification in "legally mandatory" regions like Singapore and Taiwan⁴⁰. This means Taiwan’s Anti-Scam Special Law did force Meta to do one more verification step; legal agency is not zero.

Governance Layer: Government, Law, and Technology

Facing this complex ecosystem, government responses are scattered across multiple ministries.

Ministry of Health and Welfare: Finding Account Applicants via the _Physicians Act_

Regarding the AI fake doctor case, the Ministry of Health and Welfare plans a multi-layered response: improving the digital application of medical regulations, strengthening cross-ministerial cooperation, establishing enforcement judgment standards for social media posts and AI content, planning rapid takedown mechanisms, appeal and reporting channels, and adding safety reminders for medication/emergency information⁴¹. Modeling after the US system, the Ministry plans to require Taiwanese doctor channels to disclose legal physician status in their descriptions. It will also identify account applicants via the Physicians Act and penalize them via local health bureaus, while seeking assistance from Google and Meta⁴².

Ministry of Digital Affairs: "Scam Reporting and Inquiry Network"

Former Minister of Digital Affairs Huang Yen-nan announced during his tenure that within three months, a "Scam Reporting and Inquiry Network" would be built to use AI against AI: after citizens paste suspicious messages into the platform, AI analyzes and routes them to corresponding ministries (health topics → MOHW, agriculture topics → Council of Agriculture), establishing a joint defense mechanism with e-commerce and telecom operators, and promoting real-name registration for online ads⁴³. Notably, the governance philosophy is: "Content truth/falsehood should not be judged by the government." The MOCA’s planned disinformation committee initially consists of NGO members; takedowns still require public power execution⁴³. Huang Yen-nan resigned on August 22, 2025, after his secondment period ended, returning to the Academia Sinica⁴⁴.

TikTok: Public Sector Ban, General Public Not Banned

TikTok’s legal status in Taiwan is often misunderstood. The Executive Yuan’s 2019 publication of the "Principles on Restricting the Use of Products Endangering National Cyber Security" banned TikTok, Douyin, and Xiaohongshu for public sector cyber devices and premises. However, general citizens and civil servants’ private devices are not restricted⁴⁵. The 2024 draft Regulations for the Prevention and Punishment of Fraudulent Crimes further stipulated that TikTok must comply with Chinese-invested enterprise legal norms to establish a legal representative, with the heaviest penalty being "network blocking"⁴⁶.

C2PA: Making Authentic Content Verifiable

Technically, the C2PA (Content Credentials and Provenance Alliance) protocol attempts to add an unforgeable "ID card" to digital content. Its steering committee includes Adobe, Microsoft, OpenAI, Google, Intel, BBC, and Sony⁴⁷. OpenAI automatically embedded Content Credentials in DALL-E 3⁴⁸. The Google Pixel 10 phone released on September 10, 2025, natively embeds C2PA content credentials in its camera app, achieving the highest current Assurance Level 2 of the C2PA Conformance Program, making it the first smartphone to adopt this standard at the mobile device camera level⁴⁹.

The logic of this direction is to make authentic content verifiable: shifting the burden of proof from "reader verification" to "content附带 provenance." However, C2PA has not yet popularized to short-video platforms like TikTok or YouTube Shorts, nor has it penetrated encrypted communications like LINE groups.

Taiwan’s Resistance Infrastructure

Beyond government tools, Taiwan has established a decentralized verification ecosystem: this decentralization itself is a resilience design:

The design significance of this decentralized ecosystem is checks and balances: TFC follows journalistic standards, IORG follows academic research, Cofacts follows crowd-sourcing, and LINE Official follows platform-level integration: no single institution can monopolize the definition of "what is true." This is more important than any Disinformation Prevention Special Law, because the latter will always encounter the tail-end problem of "who determines."

Intergenerational Dialogue: Where to Start

There is an anomalous observation in the dialogue layer. LINE’s user base in Taiwan exceeds 18 million, with usage rates for ages 40-49 and 50-65 both exceeding 90%⁵⁵. The elderly (65+) account for 20.25% of fraud victims⁵⁵. But multiple surveys show that when receiving suspected disinformation, about 30% of the elderly proactively verify, more active than some younger generations⁵⁶.

The stereotype "elderly = victims" needs to be corrected. Information literacy education spans the entire society; all age groups have gaps in algorithmic literacy. The AI Society Research Center at Soochow University proposed the concept of "Critical Algorithmic Literacy," defined as "Should we use it? On whose terms?": a meta-perspective higher than "digital literacy," "information literacy," or "media literacy"⁵⁷.

A concrete starting point for dialogue with the elderly can be the "Message Verification" account they are already using on LINE: LINE’s official entry integrating the four major verification platforms, paste it in to check⁵³. Cofacts’ LINE bot is also installation-free and crowd-sourced⁵¹. The commonality of these two tools is lowering the interface threshold, allowing the elderly to obtain verification results in their original social field.

International Experience: From Ukraine to the EU

Taiwan is not an isolated case. After the 2014 Crimea incident, Ukrainian journalist Dmytro Kuznetsov and students from the Kyiv Mohyla Academy of Journalism founded StopFake, which had verified over 3,000 pieces of Russian disinformation by the time the war broke out⁵⁸. After the 2022 war fully escalated, pro-Ukrainian netizens in Europe and the US spontaneously organized the "North Atlantic Fellowship (NAFO)," using satirical memes to quickly respond to Russian propaganda[^58]: this was the first appearance of purely civilian self-organized tactical countermeasures on the cognitive battlefield.

On the legislative front, Article 50 of the EU AI Act will take effect in August 2026, mandating that deepfakes must be clearly disclosed as AI-generated/manipulated, and AI-generated text on public interest topics must be labeled. Providers must ensure outputs are marked in machine-readable form⁵⁹. The European Commission released the first draft of the transparency code of practice on December 17, 2025, the second version in March 2026, and finalized it in June 2026⁶⁰.

US legislation is also advancing: the 2025 TAKE IT DOWN Act was signed into law by Trump, mandating platforms to remove non-consensual intimate images²². The DEFIANCE Act passed the Senate unanimously in January 2026 and is now in the House, granting civil rights to deepfake pornography victims²³.

Cross-border fraud has reached a new scale. On October 14, 2025, the US and UK simultaneously sanctioned Cambodia’s Prince Group, seizing Bitcoin assets worth $15 billion (the largest in US/UK history), while Taiwan seized 26 luxury cars (worth 400 million NTD). The sanction list included 9 Taiwanese companies and 3 Taiwanese individuals. The group’s office was located in Taipei 101 and Heping Grand Hotel⁶¹. This case connects AI deepfake tools, cross-border fraud, and Taiwan’s local structures into a single line.

One Structure, Three Distances

Wang said: "A group of accounts that always log on together for warmth might just really be a group of good friends." This sentence must be read together with Shen Pei-yang’s "80% has nothing to do with truth or falsehood, it is narrative attacks." The former reminds researchers and the government not to treat "synchronization" as the only accusation. The latter reminds readers not to treat "fact-checking" as the only defense line. Together, the two sentences outline the most difficult two sides of cognitive warfare: vigilance, but also restraint.

Back to the structure. An operable answer is composed of three distances:

National Level: The government’s task is to set the bottom line, not to define truth or falsehood. The MOHW finding account applicants via the Physicians Act, the MOCA building the scam reporting network, the Anti-Scam Special Law forcing Meta to do one more verification: these are structural actions at the legal level. The government must avoid becoming the "truth judge" itself, because whoever is in power will have their own truth. The lesson of the Digital Intermediary Act’s withdrawal remains.

Community Level: TFC, IORG, Doublethink Lab, Cofacts, MyGoPen, Black Bear Academy, INDSR, plus LINE Official Message Verification, this is checks and balances. Eight organizations each following their own standards, supervising and complementing each other; no single institution being caught is fatal.

Individual Level: Each person’s 15-second action. Next time you receive a "doctor’s advice" or "official subsidy" video in a LINE group, spend 15 seconds clicking into TFC or MyGoPen to search keywords, seeing if it has been fact-checked. Or forward it to the LINE Official Message Verification account for the bot to check. This action has low cost, but amplified by thousands or millions, it can suppress the spread speed of disinformation.

An elderly person sharing an AI fake doctor video is because they care about their family’s health. A young person sharing the "Unicorn So Loud" edit is because they truly resent Taiwan’s prices. Cognitive warfare precisely exploits these irreplaceable genuine emotions. Then, using false authority or exaggerated conclusions, it transforms emotion into cognitive bias in a specific direction.

"Dr. Chen Chih-ming" has disappeared from YouTube. Ba Jiong and Chen Po-yuan are still filming new documentaries, Shen Pei-yang is still meeting in the Legislative Yuan, TFC fact-checkers, Cofacts volunteers, Black Bear Academy instructors, IORG researchers: tomorrow they will still be in place.

The ultimate battlefield of cognitive warfare is trust relationships themselves. And trust relationships are not in Douyin, not in Facebook ad backends, not in CCTV’s 8-minute exposés. Trust relationships are in that action where you are willing to spend 15 seconds searching keywords, then turning back to tell your family, "I checked, this is fake."

Further Reading

• The Mountain Makers: A Century’s Bet — Hsiao Ju-chen’s 2025 documentary, five years of interviews with 80+ semiconductor pioneers, entering three CHIPS Act investment hubs in 2026: Purdue, Wisconsin, Michigan

• Threads in Taiwan — Taiwan’s social media migration history and the platform structure of the information battlefield

• Miin — Du Yi-jin and Taiwan’s AI Lab’s anti-disinformation platform, using AI to catch coordinated operation accounts (not checking content item by item)

• Open Culture Foundation — The backend organization hosting Cofacts’ verification bot, guarding Taiwan’s internet freedom

• History of Taiwan’s Online Community Migration — Understanding the role of platforms like PTT, Dcard, and Threads in cognitive warfare

• Taiwan Media and Press Freedom — News ecology and platform responsibility in cognitive warfare

• Shen Pei-yang — One of the main researchers on cognitive warfare; in 2025, became the first Taiwanese elected official investigated by China for the "crime of splitting the country"

• Black Bear Academy — A civil organization putting cognitive warfare recognition into basic civil defense camps, teaching ordinary people to maintain judgment amidst disinformation and psychological warfare

• Toxic Potatoes: Beyond 200 ppm, There Are 30 ppm, 14 Days, and a 15-Year Food Safety Scar — Anatomy of how the 2026-04 Taiwan Affairs Office "letter of allegiance" narrative precisely stepped on the 15-year food safety scar since the 2011 melamine scandal

• Taiwan and Eswatini — Concrete scenarios of Chinese language operations like "rat," "stowaway-style diplomatic breakthrough," "international laughingstock": Lai Ching-te’s reaction during the 2026-05 visit to Eswatini

• Taiwan and Paraguay — How Taiwan uses long-term cooperation to withstand Chinese market temptation and political lobbying, supporting the only diplomatic ally in South America

References