A real problem, a familiar tool
On July 2, 2026, Japan's National Police Agency (NPA) opened a public comment period — running July 3 to July 16 — on a proposed revision to the operational guidelines of the Internet Hotline Center (IHC), the body the NPA commissions to field public reports of illegal online content and request its removal from site operators and platforms (NPA notice). The revision would add two new categories to the list of content the IHC can flag for takedown: advertisements that impersonate public figures to promote fraudulent investment schemes, and posts soliciting so-called "okurikin" transfers — recruiting people to move scam proceeds through their own bank accounts for a fee. Revised guidelines are expected as early as August, following an expert-panel review (MLex).
The scale of the underlying problem is not in dispute. NPA data show SNS-based investment fraud losses hit ¥127.5 billion across 9,538 recognized cases in 2025 — up 46.3% and 48.7% respectively over 2024 (NPA fraud statistics). Provisional figures for the first two months of 2026 already show ¥28.9 billion in losses, suggesting the trend is accelerating rather than plateauing (NPA provisional data). Much of this fraud now originates in banner ads and direct messages on Instagram, YouTube and Facebook that use deepfaked or unauthorized images of celebrities and financial commentators to lend scam pitches false credibility — a genuinely novel harm that older fraud statutes were not written to address, and one where speed of takedown plausibly matters more than in most content-moderation disputes, since every day an impersonation ad stays live is another day it can find new victims.
Why the mechanism deserves more scrutiny than the goal
The steelman case for this expansion is strong: fraud is not protected expression, victims are real and often elderly or financially unsophisticated, and a reporting-and-removal channel that already works — the IHC has operated since 2006 and covers categories from child exploitation material to phishing sites and firearms solicitation (Internet Hotline Center) — is a lower-friction fix than waiting for new criminal statutes or slow-moving platform enforcement. Nobody sensible defends a Facebook ad using a deepfaked finance personality to solicit crypto "investments."
The concern is not that fraud-linked posts should be untouchable. It is that Japan is expanding an administrative takedown regime by guideline revision, not by new legislation, and doing so repeatedly.
This is the third such expansion in roughly six months. The IHC's operating guidelines were revised in February 2026 to designate crime-recruitment ads (Japan's "dark part-time job" postings) as illegal information, and the NPA is separately preparing to add content promoting illegal online gambling once the revised Basic Act on Gambling Addiction Countermeasures takes effect in September. Each individual expansion is defensible on its own facts. Cumulatively, they show a pattern: the definition of "illegal information" subject to police-initiated takedown requests is being widened through administrative guideline amendments and public-comment cycles, not through Diet debate on new offenses. The IHC's requests are formally non-binding — it asks site operators and ISPs to remove content, and reports non-compliance to police — but the practical effect on a platform weighing legal risk against a government request is closer to compliance-by-default than to a genuine choice.
The two new categories here are also less mechanically clear-cut than, say, child sexual abuse material or an unlicensed pharmacy site. "Impersonation to promote investment scams" requires a judgment call about whether an ad is fraudulent, and "solicitation of money-transfer conduct" could, in an overzealous read, sweep in legitimate discussion of remittances, crypto peer-to-peer transfers, or financial commentary that merely resembles scam patterns. Japan has no equivalent of a judicial order requirement before an IHC removal request goes out, and the public comment window — 14 days, over a national holiday stretch — is thin for a change that touches the boundary between fraud enforcement and financial speech.
What should follow
None of this argues against tackling a ¥127 billion fraud epidemic. It argues for building the accountability layer alongside the takedown power: a published error/appeal rate for IHC requests, an explicit and narrow definition of "impersonation ad" that excludes satire, criticism and legitimate financial commentary, and — ideally — a sunset or legislative ratification requirement so that guideline-by-guideline mandate creep doesn't become the default way Japan expands takedown authority. Proportionate regulation means matching the tool to the harm, not just adding categories because the mechanism already exists and comments came back short.