Taiwan's Ministry of the Interior blocked the Chinese social platform Xiaohongshu (RedNote) on December 4, 2025, ordering domestic internet providers to stop resolving its domain for one year. Nine months later, Interior Minister Liu Shih-fang has signaled the block will likely be renewed: "It may be blocked for another year," she said, adding that the company remains welcome to send a written response and "come to Taiwan to explain" — an offer Xiaohongshu's Shanghai-based parent, Xingyin Information Technology, has not taken up since Taipei first requested an improvement plan through the Straits Exchange Foundation in October 2024.
A Fraud Statute, Applied Narrowly
The legal basis matters. The block was issued under Article 42 of the Fraud Crime Prevention and Control Act, which lets a competent authority order an ISP to "stop resolving or restrict access" to a site or app when necessary to prevent the public from encountering fraud. It is not a general content-control or disinformation law, and Taiwan's government has been careful to frame the action in those terms: the Interior Ministry's press release cites 1,706 fraud cases tied to the app since 2024, with losses exceeding NT$247.68 million (~US$7.9 million), and says National Security Bureau testing found Xiaohongshu met zero of Taiwan's 15 baseline cybersecurity indicators.
The steelman for the block is genuinely strong. Xiaohongshu operates in Taiwan with more than 3 million users and no local office, no legal representative, and — per the ministry — no functioning channel for law enforcement to request data in active fraud investigations. Meta, Google, LINE and TikTok all maintain Taiwan-facing compliance functions; Xiaohongshu, by the government's account, has simply ignored requests for nearly two years. A platform that won't establish even a nominal point of contact isn't being singled out for its politics — it's failing a threshold every other major foreign platform operating in Taiwan has cleared. Since the block, Xiaohongshu-linked fraud cases have reportedly fallen roughly 80% and associated losses around 55%, according to Radio Taiwan International's reporting on the ministry's figures.
Where the Justification Runs Out
But a fraud-prevention statute is being used to sustain what is functionally becoming an indefinite platform ban, and that is a different thing than a one-time emergency measure. Article 42 was written for urgent, time-bound intervention — "immediate necessity" is the operative phrase in its text. Nothing in the statute or in the ministry's public statements specifies what Xiaohongshu would actually need to do to get the block lifted: no published compliance checklist, no third-party audit standard, no sunset test beyond an open-ended annual renewal tied to the company's silence. A one-year block that quietly becomes a two-year block on the same emergency footing, with the burden of proof entirely on the company to first request a meeting, starts to look like policy by inertia rather than a proportionate, reviewable enforcement action. Deputy Interior Minister Ma Shih-yuan said in December 2025 that Xiaohongshu "must submit improvement plans to strengthen anti-fraud measures and cybersecurity" before restrictions are lifted — a reasonable condition in principle, but one Taipei has not made concrete or verifiable from the outside.
That ambiguity is also where free-expression and precedent concerns creep in, separate from the fraud rationale. A DNS-level block doesn't distinguish fraudulent listings from the ordinary posts, product reviews and cultural content that make up the bulk of what Xiaohongshu's roughly 3 million Taiwanese users actually consume — it degrades the whole platform indiscriminately, and legal commentators have already begun flagging the free-expression trade-off in that design choice. A tool built for shutting down phishing domains is being stretched to govern an entire foreign social network indefinitely, without the narrower, app-specific remedies — geofenced ad restrictions, mandatory local data-request channels, in-app fraud warnings — that address the same harm with less collateral reach.
The Same Government Has a Better Model Next Door
What makes the contrast sharper is that Taiwan's own Ministry of Education just demonstrated the proportionate alternative. On August 28, 2026, MOE released a free lesson plan — compiled by National Defense University journalism professor Fu Wen-cheng — that uses Xiaohongshu as a case study for teaching students to spot algorithmically targeted scam content and protect their data, distributed through the ministry's Media Literacy Educational Resources Network. Crucially, the curriculum does not ask students to delete the app or stop using it; it teaches critical thinking about how the algorithm and scam tactics work, while leaving the choice to use the platform with users.
That is the more defensible model for platform governance generally: transparency, media literacy, and targeted enforcement against specific fraudulent conduct, rather than a blanket access block renewed on an undefined timeline. Taiwan does not need to choose between tolerating a genuinely non-compliant foreign platform and blocking it indefinitely. It could publish the specific, auditable conditions Xiaohongshu must meet, commit to lifting the block automatically once they're verified, and let the Education Ministry's literacy-first approach — already running in parallel — do more of the work the DNS block is currently doing badly.