A First Test of Japan's Platform Law
On 24 July 2026, Japan's Ministry of Internal Affairs and Communications (MIC) issued its first-ever corrective recommendations under Article 28 of the Information Distribution Platform Countermeasures Act (情報流通プラットフォーム対処法), the reworked and renamed Provider Liability Limitation Act that took effect in April 2025. The targets were Google, X, Meta, Dwango (operator of the video platform Niconico), and Shonan Seibu Home (which runs the anonymous bulletin board bakusai.com) — the five "large-scale specific telecommunications service providers" MIC designated under the law in April 2025 (soumu.go.jp).
The finding: each company's fiscal-2025 transparency disclosure — the annual report on how it receives, evaluates, and acts on removal requests for defamatory or otherwise rights-infringing content — omitted required items. Google's report was missing 22, X's 18, Shonan Seibu Home's 16, Meta's 7, and Dwango's 4 (ITmedia NEWS). All five must republish corrected disclosures by 31 August 2026; Google and X, whose gaps were largest, must also submit improvement plans and file quarterly progress reports through April 2027 (Anadolu Agency).
The Case for the Rule
The steelman here is straightforward and worth taking seriously. Article 28 doesn't ask platforms to remove anything or adopt any particular moderation policy — it asks them to publish, once a year, how their existing removal process actually worked: how many requests came in, how many were acted on, how notice was given to affected users, and how the platform graded its own performance. That is a modest, content-neutral ask, and it exists because opaque moderation has a real victim class in Japan: individuals who report defamatory posts and have no visibility into whether — or why — nothing happened. Requiring platforms to show their work is a lighter-touch alternative to the takedown mandates and speech-defining rules other jurisdictions have reached for, and MIC's own escalation ladder reflects that restraint: a recommendation first, a formal corrective order only if the recommendation is ignored, and criminal penalties — up to one year's imprisonment or a ¥1 million fine — reserved for defying an order, not for the original disclosure gap (Monolith Law).
Why the Gradient Still Matters
That said, the scale of what counts as "incomplete" here should give pause before treating this as evidence of bad faith. Google's disclosure was missing 22 of a lengthy list of required data points — a compliance-mapping failure against a still-new ministerial ordinance, not a refusal to report anything. Complex, multinational transparency regimes reliably produce this kind of gap in their first enforcement cycle: the EU's Digital Services Act transparency-reporting requirements went through several rounds of Commission guidance before major platforms' reports were considered adequately complete, and the DSA's own template evolved after initial submissions from very large platforms fell short. Treating a first-cycle disclosure shortfall as equivalent to defying a takedown order would blur a distinction the law itself is careful to preserve.
MIC's response so far respects that distinction. Nobody has been fined. Nothing has been ordered removed. The remedy is republication with corrections, plus, for the two companies with the largest gaps, a structured improvement plan and follow-up reporting rather than an immediate penalty. That is proportionate regulation working as designed — using the recommendation-before-order structure to give platforms a real chance to fix a paperwork failure before it becomes a legal one.
The Risk to Watch
The genuine test is what MIC does if the 31 August republication is judged inadequate, and how it treats Shonan Seibu Home, an anonymous-posting board with far less compliance infrastructure than Google or Meta, differently from the two U.S. giants it singled out for improvement plans. A regulator that escalates quickly and proportionately toward the smaller, less-resourced operator while giving Google and X open-ended runway would undercut the law's claim to even-handedness. Equally, if MIC starts treating disclosure completeness as a lever to pressure substantive moderation choices — rather than sticking to the neutral question of "did you report what you were asked to report" — the light-touch model this case currently represents would tip into something closer to a backdoor content mandate.
For now, the record supports treating this as a working example of transparency regulation done right: specific, content-neutral, proportionate, and reversible if platforms comply by the deadline. Other jurisdictions weighing platform transparency rules — and observers assessing whether Japan's model deserves emulation — should watch the 31 August republication and MIC's next move as the real signal, not this first recommendation itself.