Japan's communications ministry has opened a public comment period on a draft guideline that tells large online platforms how to curb election misinformation, and the draft is notable for what it declines to compel. Comments are accepted from 24 October to 24 November 2026, and the draft guideline is meant to take effect with the underlying law on 1 March 2027.
The law behind it
The guideline implements Law No. 58 of 2026, a member-sponsored bill the Diet passed on 13 July 2026. It amends the Public Offices Election Act and the Information Distribution Platform Act. Under the amended platform act, a large operator designated by the communications minister must take measures, suited to its service, against the spread of information that threatens election fairness: content that breaks the law, false information and information that distorts the facts. It must also publish how it did once a year.
The draft does not name the designated operators.
Recommendations, not a checklist
The draft says its measures are examples. Operators are not required to adopt all of them, and the specific choices are left to each operator's own judgment. Because Diet deliberations settled on maximum respect for free expression, the draft says, the content of an operator's measures is not subject to administrative action or penalties, and the law does not oblige platforms to rule on individual posts. The draft also says it does not ask operators to delete individual items, and it warns that suspending content without need, when other options exist, would restrict expression excessively.
The recommended menu is long:
- Extra staff for election periods, including legal specialists.
- Removal of posts and account suspensions under platform terms of service.
- Tighter eligibility checks for creators who are paid by view counts, a visible marker for paid creators, and suspension of payments when election-fairness rules are breached.
- Rules and labelling functions for AI-generated content, plus detection and warnings even where the poster has not labelled it.
- An "identity verified" badge and a display of the poster's country or region.
- Prominent placement for reliable sources such as election commissions, candidates and parties, and lower ranking for information that threatens election fairness.
- Detection of bulk or automated posting.
- Cooperation with fact-checkers and outside researchers.
The draft also says operators should publish their criteria in advance and tell affected users why a measure was applied.
Where the penalties sit
The enforceable obligation is the annual public report on measures taken. The draft recounts the chain: the minister may require reports, and a failure to report or a false report carries a fine of up to ¥500,000. The minister may recommend corrections and, if an operator ignores a recommendation without good reason, order them. Breaching an order is punishable by up to one year in prison or a fine of up to ¥1mn, and a company can face a fine of up to ¥100mn.
The draft asks that results be reported quantitatively where possible and that operators evaluate themselves as critically and objectively as they can. The reporting format is to be fixed in a ministerial ordinance amending the platform act's enforcement regulations, which is the second item in the comment call.
Separately, the draft restates duties for people posting during campaigns. Anyone distributing AI-made or AI-altered images in campaign material online between the election notice and polling day must label them, unless the alteration is minor by common standards or the image could not be mistaken for an actual photograph. Publishing false statements about a candidate with intent to defeat them can bring up to four years in prison or a fine of up to ¥1mn.
How to respond
Submissions go through e-Gov, by email or by post, must arrive by the deadline and must be written in Japanese. Comments over 1,000 characters need a summary. After the comment period the ministry plans to finalise the provisions in light of the responses.