Japan

Japan has a complication the other markets in this catalogue do not, and it is worth meeting before any of the rules: the English you are reading has no legal effect. The government publishes English texts and labels them plainly as reference material. An institution that builds a control from the English alone has built it from a document that is, by its own statement, not the law.

Past that, two of Japan's choices catch teams who expect the European pattern — the vocabulary is its own, and the deadline you are looking for is not in the Act.

The English text is reference material, not law

The Personal Information Protection Commission publishes an English text of the Act and attaches this note: "Only the original Japanese texts of the laws and regulations have legal effect, and the translations are to be used solely as reference materials to aid in the understanding of Japanese laws and regulations."

The Financial Services Agency does the same with its AML guidelines, whose first page is headed "[Provisional Translation]" and says "The original texts of the Guidelines are prepared in Japanese, and this translation is only provisional."

So the practical rule for a compliance function operating in Japan is that English is for understanding and Japanese is for deciding. Any control, policy or system rule that turns on exact wording needs to be settled against the Japanese text by somebody who reads it.

Bears on: AI, data protection and privacy, Cyber security foundations, eKYC and digital identity verification, IT service management foundations

The vocabulary is its own

Japan does not say controller or data subject. The Act on the Protection of Personal Information (Act No. 57 of 2003, English text reflecting amendments to Act No. 37 of 2021) says "businesses handling personal information" for the organisation and "the identifiable person" for the individual. The regulator is the Personal Information Protection Commission.

A document written in European vocabulary reads here as though it were drafted for elsewhere, which is the first thing a reviewer notices.

Bears on: AI, data protection and privacy, Cyber security foundations, eKYC and digital identity verification

A leak is reported, and the deadline is not in the Act

Article 26 is headed Reporting Leaks. Paragraph (1) requires a business handling personal information to report "leaks, loss or damage and other situations concerning the security of the personal data they handle, which are prescribed by Order of the Personal Information Protection Commission as those likely to harm individual rights and interests", to the Commission. Paragraph (2) then requires it to "notify the identifiable person of the occurrence of the situation", unless that is difficult and "the necessary alternative measures are taken to protect that person's rights and interests".

Two things follow. The Act itself fixes no number of hours — both the reportable situations and the manner of reporting are set by Commission Order, so the deadline has to be looked up there rather than read off the statute. And Article 26(1) carries a proviso that matters for outsourcing: where another business or an administrative entity has entrusted the handling of the data, and the handler has notified that entrusting party, the handler does not also report to the Commission.

If your incident runbook says "72 hours" and names no country, it is wrong in Japan — not because the period is different, but because the statute is not where the period lives.

Bears on: Cyber security foundations, Incident response and post-incident review, IT service management foundations, Third-party cyber risk

There is no right against a decision made without a person

The words "automated" and "profiling" do not appear in the Act at all. There is no right not to be subject to a decision based solely on automated processing, and no right to have such a decision reconsidered.

Read that beside India, where the position is the same, and the pattern becomes the point: the European right under Article 22 of the General Data Protection Regulation is the provision most national laws on this subject are modelled on, but it is not the norm across Asia. A design that assumes a human reviewer will be legally required does not travel. If a person belongs in the loop, put them there because the risk warrants it, not because you expect the law to insist.

Bears on: AI agents: the basics, The role of AI in fraud, AI governance foundations, Fraud models in production, eKYC and digital identity verification, Agentic AI: how agents work, Agentic workflows and multi-agent systems, Testing, evaluating and monitoring AI agents

AML sits on two Acts, and the guidelines set a minimum rather than a ceiling

The FSA's guidelines state where the duties come from. "The basic requirements on anti-money laundering and combating the financing of terrorism ('AML/CFT') in Japan, such as identification and verification at the time of transactions, are prescribed in the Act on Prevention of Transfer of Criminal Proceeds ('Criminal Proceeds Act', or APTCP), the Foreign Exchange and Foreign Trade Act ('Foreign Exchange Act'), and other relevant laws and regulations."

On the risk-based approach the guidelines are unusually direct. It "is established as a central principle of the Financial Action Task Force ('FATF') Recommendations", and "the risk-based approach is a minimum standard that financial institutions participating in Japan's financial system should implement."

The guidelines are also structured in a way worth copying into your own documentation. Each section states "Required actions for a financial institution", followed by "A financial institution shall:" and a numbered list. A supervisor reading your framework can line it up against that list directly, which is an argument for writing your own controls in the same shape.

Bears on: AML screening foundations, Transaction monitoring and reporting, Business-wide risk assessment, Auditing monitoring and screening

Customer due diligence is defined wider than checking a customer

The FSA's guidelines define the term for themselves, and the definition is broader than the per-customer reading most teams carry. CDD "refers to, among other mitigation measures, a series of processes in which a financial institution collects and verifies information about specific customers and their activities and transactions in light of the results of its own risk assessment, compares that information with the results of the risk assessment, and determines and implements effective measures necessary to mitigate those identified risks."

Note what is doing the work there: the institution's own risk assessment, twice. Due diligence here is not a fixed list of checks applied to everyone; it is whatever the assessment says is needed, and the assessment is the thing a supervisor will ask to see first.

Bears on: Customer due diligence and KYC, Transaction monitoring and reporting, Auditing monitoring and screening

Sources

The official documents this course relies on. Laws and guidance change, so check the current version.

  1. Act on the Protection of Personal Information (Act No. 57 of 2003), English text · Japanese Law Translation, Ministry of Justice, Japan
  2. Laws and Policies, including the consolidated English text of the Act · Personal Information Protection Commission, Japan
  3. Guidelines for Anti-Money Laundering and Combating the Financing of Terrorism (provisional translation) · Financial Services Agency, Japan
Contact · Saolix GroupPeople. Products. Possibilities.

Your next chapter.
Let’s build it together.

Explore a product, discuss a project, or find the right expertise. Tell us where you want to go. We’ll help you take the next step.

Compliance & agentic AIManaged servicesConsulting & training