Taiwan heads into its November 2026 local elections with no binding anti-disinformation platform law — and, on closer inspection, that absence is less a governance failure than a deliberate, defensible choice. A widely-discussed April 10, 2026 analysis in Just Security, "The Tightrope Walk of Democratic Defense," frames Taiwan's predicament as a "governance trilemma": institutional, political, and economic constraints that block any EU-style omnibus content statute even as Beijing's information operations grow more sophisticated. The piece treats the gap as a vulnerability. We think it is closer to a calculated trade-off — and the right one.
The strongest case for a binding law
The argument for a comprehensive platform statute deserves a fair hearing, because it is not frivolous. China's cognitive-warfare capability is escalating in kind, not just degree. The leaked "GoLaxy papers" — roughly 400 internal pages from a Beijing firm tied to the Chinese Academy of Sciences, obtained by Vanderbilt University researchers — describe AI persona armies built to profile politicians, detect narrative shifts, and flood platforms with tailored, human-seeming propaganda. Against an adversary that can spin up thousands of synthetic accounts overnight, Taiwan's existing tools look outmatched: election-law penalties for deepfakes are slow, backward-looking, and cannot reach the upstream state orchestrating the campaign. A standing legal mechanism that compelled platform transparency and rapid response, the argument runs, would let democracy keep pace. That is a serious concern, and dismissing it would be a strawman.
Why Taiwan keeps saying no
But Taiwan has already run this experiment and watched it fail on democratic terms. The National Communications Commission's 2022 Digital Intermediary Services Act (DISA), modeled on the EU's Digital Services Act, was withdrawn within months. As Just Security notes, opposition parties "equated its emergency content-restriction provisions with state censorship." The public went further: more than 30,000 people voted against the draft on the government's own online policy-participation platform, against fewer than 150 in support. Citizens circulated mock notices reading "this content has been removed for violating the Digital Intermediary Services Act." In a society that lived under four decades of martial-law information control, any statute that hands the state a content kill-switch — however well-intentioned — triggers a democratic immune response. The NCC shelved the bill before it ever reached the legislature.
That reflex is not paranoia; it is the correct instinct. A broad takedown power justified today by anti-PRC necessity becomes, under a different government, a tool aimed at domestic dissent. The 30,000 objections were a society pricing in that risk accurately.
The targeted alternative is already working
What Taiwan built instead is a sector-specific patchwork — and it is more functional than the "trilemma" framing admits. The Fraud Crime Hazard Prevention Act, in force since July 31, 2024, imposes concrete obligations on online advertising platforms: verify advertiser identities, remove fraudulent ads within 24 hours of notice, and bear joint liability for harm they fail to act on. According to the Ministry of Digital Affairs, six services from four companies — Google, YouTube, LINE, Facebook, Instagram, and TikTok — are now designated operators, with subsidiary regulations effective November 30, 2024.
Crucially, this regime is tied to a defined harm — financial fraud — not to the contested category of "disinformation." It regulates conduct and money flows rather than viewpoints. That is the proportionality the DSA model lacked. When Taiwan extended the same statute to block RedNote (Xiaohongshu) for one year in December 2025, the Taipei Times reported the basis as 1,706 documented fraud cases and a National Security Bureau review in which the app failed all 15 cybersecurity indicators. Radio Taiwan International put the losses at NT$247.7 million (about US$7.9 million) and quoted NCC acting chair Chen Chung-shu locating the core risk in China's "draconian information-sharing laws" and the platform's lack of any local legal presence to receive complaints.
The honest gap — and how to fill it
The critics are right about one thing: none of this directly answers foreign information manipulation and interference (FIMI). The anti-fraud law was never designed to, and the RedNote ban is narrow and easily circumvented by VPN. There is a real coverage hole between "financial fraud" and "hostile state propaganda."
The answer, though, is not to resurrect emergency takedown powers. It is to close the gap with the lightest-touch tools that match the threat: mandatory transparency on coordinated inauthentic behavior and political ad funding, algorithmic disclosure, and continued state support for Taiwan's world-leading civil-society fact-checkers and the participatory, bottom-up rebuttal networks that Just Security itself credits. These attack the delivery mechanism of influence operations — inauthenticity and opacity — without empowering any official to decide what is true. Against GoLaxy-style synthetic personas, an authenticity-and-disclosure mandate is more surgical than a content ban, and far harder to abuse.
Taiwan's slow, contested, sector-by-sector path looks messy next to a single comprehensive code. But messy and rights-protective beats tidy and dangerous. A democracy that defends itself by acquiring the censor's toolkit has already conceded part of what it set out to protect. Taipei's restraint is not the absence of a strategy — it is the strategy.