Jurisdiction Baseline · AI × Policy

AI Regulation in Japan.

At a glance

AI-Specific Law

In force; promotes AI, without broad licensing

Privacy Rules

APPI governs personal data used throughout AI

Training and outputs may require permission

Consumer Protection

Misleading AI claims can trigger corrective action

Japan has an AI-specific law, the Act on Promotion of Research and Development and Utilization of AI-related Technology, which fully took effect on September 1, 2025. It mainly promotes AI development and sets governance responsibilities rather than imposing EU AI Act-style licensing or risk categories, so companies are primarily regulated through personal data protection, copyright, consumer and advertising, and sector-specific laws, with AI enforcement still largely guidance-led.

In force today5 instruments

Act on Promotion of Research and Development and Utilization of AI-related Technology in forceAI-specific

Japan's first AI-specific law establishes national AI governance structures, an AI Basic Plan, and government measures to promote trustworthy AI. It does not create a general licensing system, prohibited-practice list, or broad mandatory conformity assessment for private AI providers, but it supports government guidance, information requests, recommendations, and public disclosure in cases of serious risk or non-cooperation.

Read the full text ↗ · 2025

AI / automated decisionsSafety & security

Act on the Protection of Personal Information in force

The APPI governs the collection, use, retention, security, transfer, and disclosure of personal information used to train, test, or operate AI systems. Companies must define purposes of use, handle sensitive personal information carefully, manage vendors and overseas transfers, respond to data-subject rights, and report certain breaches; the PPC has issued specific warnings on generative AI use.

Read the full text ↗ · 2003

Data processingDisclosureSafety & security

Unlock the full baseline

Get the remaining 3 instruments and the founder FAQ. Free: subscribe once and every briefing on this device unlocks.

Already subscribed? Unlock →

Copyright law applies to training data, retrieval corpora, prompts, model outputs, and downstream distribution. Japanese law contains exceptions that can permit certain information-analysis uses, but permission may still be required where uses exceed the exception, substitute for exploitation of the work, or involve infringing outputs; the Agency for Cultural Affairs' AI copyright interpretation is useful but expressly nonbinding.

Read the full text ↗ · 1970

Content moderation

Act against Unjustifiable Premiums and Misleading Representations in force

AI-generated claims about product performance, accuracy, safety, savings, or capabilities must not materially mislead consumers. The Consumer Affairs Agency can require reasonable substantiation for performance claims and impose corrective measures or surcharges, so AI vendors should validate marketing claims and disclose material limitations where needed.

Read the full text ↗ · 1962

Consumer protectionDisclosure

Act on Improving Transparency and Fairness of Specified Digital Platforms in force

This regime matters mainly to designated large digital platform operators, including certain online marketplaces and app-store businesses that use or distribute AI systems. Covered operators must disclose terms and procedures, maintain fair dealing systems, and submit annual reports, but ordinary AI SaaS providers are generally outside its scope.

Read the full text ↗ · 2020

DisclosureConsumer protection

Also worth noting

The July 17, 2026 amendments to the Act on the Protection of Personal Information were promulgated but are mostly scheduled to apply on a future date within two years, with implementing orders, PPC rules, and guidelines still to come.
The Cabinet adopted an AI Basic Plan on December 23, 2025 and a revised plan on July 14, 2026. Monitor follow-on measures under the AI Act, including the AI guideline framework and any move from voluntary governance toward more specific duties.
METI and MIC's AI Guidelines for Business Version 1.2, updated April 1, 2026, remain nonbinding but are the main practical benchmark for risk management, transparency, security, human oversight, and incident response.

Pending and recent legislation2 measures

JP 第221回閣法53号 Diet Bill No. 53 of the 221st Session: Bill to Partially Amend the Act on Promotion of Administration Using Information and Communications T

Passed · 2026-07-17

The bill updates Japan’s legal framework for ICT-enabled public administration and information-processing policy, including measures supporting the government’s promotion and use of artificial-intelligence-related technologies. It also expands the institutional basis for public-s

JP 第221回閣法54号 Bill No. 54 of the 221st Diet: Bill to Partially Amend the Act on the Protection of Personal Information and Other Acts

Passed · 2026-07-17

The bill amends Japan’s personal-information protection framework, including rules governing the use, provision, and protection of personal data relevant to data-intensive technologies such as AI. It also revises related institutional and enforcement provisions.

Live from the Global AI Policy Radar.

Founder FAQAI law, regulators, what applies, what is next, penalties

Is there an AI-specific law in force here?

Yes. Japan’s Act on Promotion of Research and Development and Utilization of AI-related Technology fully took effect on September 1, 2025. It is mainly a promotion, governance, and coordination framework, not an EU AI Act-style licensing or risk-classification regime, and most private-sector duties are responsibilities or efforts supported by government guidance.

Who are the key regulators for AI?

The Personal Information Protection Commission oversees personal information issues, including data used to train, test, or operate AI systems. The Consumer Affairs Agency handles misleading consumer and advertising claims. The government also uses the AI-specific Act to coordinate policy, issue guidance, request information, make recommendations, and publicly disclose serious risks or non-cooperation.

Which rules apply to AI systems today?

AI deployments are principally governed by the Act on the Protection of Personal Information, copyright law, consumer and advertising rules, and applicable sector-specific rules. The AI-specific Act also applies, but it does not create a general licensing system, prohibited-practice list, or broad mandatory conformity assessment for private AI providers. Large designated digital platforms may also face transparency and fairness obligations.

What is coming next, and when?

The AI-specific Act is already fully in force as of September 1, 2025. It establishes national AI governance structures and an AI Basic Plan, with further government guidance and measures to promote trustworthy AI. The source does not provide additional dates for future rules.

What are the enforcement and penalty risks?

AI-specific enforcement remains comparatively light and guidance-led, but the government may request information, issue recommendations, and publicly disclose serious risks or non-cooperation. More concrete risks arise under adjacent regimes, where the Personal Information Protection Commission and Consumer Affairs Agency can investigate and take administrative action. Misleading performance claims may lead to corrective measures or surcharges, while privacy breaches, improper data handling, copyright infringement, and failures by covered platforms create additional exposure.

Not legal advice. For educational purposes only. AI-researched against official sources (2026-08-06), links verified.

Need a company-specific assessment? Explore Nomos →