Taiwan's Unique Disinformation Threat
For eleven consecutive years, Taiwan has ranked as the country most affected by disinformation, according to the Varieties of Democracy (V-Dem) project. A 2024 survey by Truth Matters found that 95 percent of Taiwanese respondents had encountered misinformation. In 2025 alone, the National Security Bureau (NSB) identified more than 2.314 million pieces of Chinese disinformation circulating across social media, distributed through over 45,000 fake accounts operated by firms including Haixunshe and Huya under direction from Beijing's Central Publicity Department.
This is not the ambient noise of democratic discourse. It is coordinated cognitive warfare, designed — per the NSB's January 2026 assessment — to "deepen internal divisions within Taiwan, weaken Taiwanese people's will to resist, and win support among Taiwanese people for the PRC." Given that, the argument for forceful government intervention is serious: civil society fact-checkers, however sophisticated, cannot match the operational tempo of a state-backed information campaign running at machine scale. Defenders of stronger platform regulation are not wrong about the threat.
The question is whether Taiwan's current legislative architecture is actually up to the task.
The DISA's Brief Life and What It Revealed
In June 2022, the National Communications Commission (NCC) proposed the Digital Intermediary Services Act (DISA), modeled on the EU's Digital Services Act. It would have imposed tiered content-accountability obligations on large platforms — transparency requirements, expedited takedown mechanisms, and regulatory audit rights. The bill never reached parliament. Within weeks, more than 30,000 citizens voted against it on the government's public participation platform; fewer than 150 expressed support.
The reaction was not anti-regulation sentiment in the abstract. Critics — including civil society groups otherwise hawkish on Chinese disinformation — objected specifically to vesting broad content-removal authority in the executive branch. In a democracy that institutionalized civilian rule within living memory, that concern carries real weight. The lesson from the DISA's collapse is not that Taiwan rejects platform accountability; it is that the legitimacy of the mechanism matters as much as the goal. Executive-controlled takedowns without independent oversight trigger justified suspicion regardless of the stated rationale.
The Anti-Fraud Pivot
With the DISA dead, the government found a different vehicle: the Fraud Crime Hazard Prevention Act (FCHPA), passed July 2024 and fully effective from January 21, 2026. The act designated seven platform surfaces — Google Search, YouTube, LINE, Facebook, Instagram, TikTok, and Threads (added September 2025) — as regulated online advertising operators. Under Articles 31 and 32, these platforms must verify and disclose advertiser identities, label deepfake and AI-generated advertising content, and remove fraudulent content upon notification. Fines range from NT$200,000 to NT$10 million for standard violations and up to NT$100 million for serious non-compliance.
Enforcement has already begun. In May 2025, the Ministry of Digital Affairs (MODA) issued its first administrative penalty against Meta for failing to disclose advertiser identities on Facebook ads — a per-violation structure that creates meaningful cumulative exposure for repeat non-compliance. Major platforms have also integrated into the government's "Green Channel" rapid-reporting system, giving law enforcement a direct route to flag fraudulent content for removal.
The December 2025 blocking of Xiaohongshu (RedNote) showed how far the framework can reach. The Chinese social media platform — with over 3 million Taiwanese users — failed all 15 of Taiwan's cybersecurity compliance benchmarks, ignored regulatory requests across 52 days, and was linked to 1,706 fraud cases totalling NT$247 million in losses. MODA ordered a one-year DNS block under Article 42 of the FCHPA. The contrast with cooperative platforms is deliberate: Meta, Google, LINE, and TikTok all comply; non-cooperation now carries real consequences.
Where the Architecture Strains
The anti-fraud pivot is pragmatically clever, but it has a structural ceiling. Fraud law targets verifiably false commercial content — fake celebrity investment endorsements, deepfake product promotions, scam recruitment posts. It cannot cleanly reach the broader category of political disinformation: narratives engineered to deepen social divisions, manipulate perceptions of cross-strait relations, or suppress minority viewpoints. Those tactics rarely meet the legal threshold of "fraud" as conventionally defined under criminal or administrative law.
As researcher Hung Pu-chao of Tunghai University's Cross-Strait Research Center warned in January 2026: "We cannot use laws from the analog era to fight cognitive warfare in the digital age." The Anti-Infiltration Act, in force since 2020, criminalizes certain activities tied to foreign-state political interference — but has produced only 5 convictions from 127 indictments since enactment, largely because courts default to suspended sentences absent statutory minimum penalties.
The Freedom House 2024 internet freedom report scores Taiwan 79 out of 100 — firmly in the "Free" tier — but flags a procedural concern with the anti-fraud blocking mechanism. The DNS Response Policy Zone (RPZ) emergency-blocking pathway processed 36,559 domain restrictions via administrative requests between June 2023 and May 2024, compared to just 29 ordered through courts. In April 2024, the same system accidentally blocked legitimate WordPress-hosted sites for seven hours. Speed without independent oversight creates legitimacy costs that will compound as the system scales.
The Next Move
Proposed amendments to the Anti-Infiltration Act, described by the Mainland Affairs Council in January 2026 as "urgent and necessary," would add specific provisions addressing disinformation dissemination and impose new compliance obligations on ISPs operating databases in Taiwan. Those amendments remain stalled in a fractured legislature whose opposition bloc has blocked national security legislation repeatedly.
Taiwan's civil society infrastructure — MyGoPen, Taiwan FactCheck Center, Cofacts, and DoubleThink Lab — remains among the most sophisticated disinformation-resistance ecosystems anywhere, and media literacy is now embedded in the national school curriculum as one of nine core values. These are genuine structural strengths that top-down regulation cannot substitute for.
But neither fact-checkers nor a fraud law were designed to govern the political advertising and coordinated-amplification ecosystem that shapes civic opinion at scale. The DISA failed because its mechanism lacked adequate checks — not because the goal was wrong. A narrower, purpose-built framework mandating political ad transparency, disclosure of coordinated inauthentic behavior, and genuine judicial oversight of emergency takedowns could close that gap without repeating the overreach that sank the original bill. Taiwan understands the threat more clearly than almost any other democracy. The unfinished work is building a response with both teeth and legitimacy.