Loading
Japan enacted its first national AI law in 2025, the AI Promotion Act, but chose an innovation-first model: the Act has no penalties, no prohibitions and no mandatory conformity assessment. It sets principles and creates an AI Strategy Headquarters, chaired by the Prime Minister, which adopts an AI Basic Plan. The obligations that actually bind AI come from existing law such as the personal-data law and the Copyright Act, while the AI-specific layer, including the widely followed AI Guidelines for Business, is soft law. This page sets out the instruments, the dates, and which ones are binding versus voluntary, with links to the primary sources.
Last reviewed: 22 July 2026 · A factual snapshot; AI policy is evolving, so confirm against the primary sources linked below
Japan is often described as having the most innovation-friendly AI law of any major economy, and the description is fair. The AI Promotion Act, passed in May 2025 and promulgated in June 2025, is a promotion and framework law. It states national principles, obliges the government to support AI research and use, and creates an AI Strategy Headquarters within the Cabinet, chaired by the Prime Minister, which draws up an AI Basic Plan. It does not create sanctions: there are no fines, no prohibited uses, and no mandatory conformity assessment.
So the rules that can actually be enforced against an organisation come from existing law applied to AI: the Act on the Protection of Personal Information (APPI) for personal data, the Copyright Act, whose Article 30-4 exception is central to how AI is trained in Japan, and sector rules. On top of that sits an influential soft-law layer: the METI and MIC AI Guidelines for Business (now at Version 1.2), the AI Basic Plan, and Japan's international work through the G7 Hiroshima AI Process.
A 2018 amendment added Article 30-4, permitting the use of copyrighted works for information analysis, which includes training AI models, without the rightsholder's authorisation where the purpose is not to enjoy the work's expression. It carries a proviso: it does not apply where the use would unreasonably prejudice the interests of the copyright owner.
Under its 2023 G7 presidency, Japan launched the Hiroshima AI Process, which produced international Guiding Principles and a voluntary International Code of Conduct for organisations developing advanced AI. It is soft law rather than binding regulation.
METI and the Ministry of Internal Affairs and Communications (MIC) jointly published the AI Guidelines for Business, consolidating three earlier guidelines on development, use and governance into a single voluntary framework for the responsible development, provision and use of AI.
The Diet passed the Act on Promotion of Research and Development, and Utilization of Artificial Intelligence-related Technology (Act No. 53 of 2025), Japan's first national AI law. It was promulgated on 4 June 2025. It is a promotion and framework law: it sets basic principles and creates governance machinery, but imposes no penalties, no prohibitions and no mandatory conformity assessment.
The chapters creating the AI Strategy Headquarters and the AI Basic Plan took effect. The Headquarters is a Cabinet body chaired by the Prime Minister, with all Cabinet ministers as members; it held its first meeting on 13 September 2025 and is tasked with drawing up a comprehensive AI Basic Plan.
The Cabinet formally decided Japan's first AI Basic Plan under the Act, setting mid to long-term national measures consistent with the Act's principles. It was later revised, with a second plan decided on 14 July 2026, confirming the Act works through planning and coordination rather than enforceable rules.
The Financial Services Agency updated its discussion paper on the sound use of AI in the financial sector to version 1.1. It takes a principles-based, agile stance and maps AI use cases and risks rather than imposing binding requirements.
METI and MIC published Version 1.2 of the AI Guidelines for Business, the current edition. The Guidelines are a living document and remain voluntary, but they function as the practical standard of care against which responsible AI practice in Japan is measured.
Because the AI Promotion Act carries no penalties, the distinction that matters most is whether an instrument can be enforced or is guidance an organisation is expected to follow. The table below sorts the main instruments accordingly.
| Instrument | Status | Applies to / owner |
|---|---|---|
| AI Promotion Act (Act No. 53 of 2025) | In force, no penalties | National framework; establishes the AI Strategy Headquarters and Basic Plan, with a non-binding duty on businesses to cooperate |
| Act on the Protection of Personal Information (APPI) | Binding | Any organisation using personal data, including in AI (enforced by the PPC) |
| Copyright Act, Article 30-4 | Binding | AI training on copyrighted works, permitted subject to the unreasonable-prejudice proviso |
| AI Guidelines for Business (Version 1.2, Mar 2026) | Voluntary framework | Developers, providers and users of AI (METI and MIC) |
| AI Basic Plan (Dec 2025, revised Jul 2026) | Government plan, not a rule | Whole-of-government direction under the AI Promotion Act |
| FSA AI discussion paper (v1.1, Mar 2026) | Discussion paper, not a rule | Financial institutions (Financial Services Agency) |
| PPC alert on generative AI (Jun 2023) | Guidance | Organisations using generative AI with personal data |
| G7 Hiroshima Code of Conduct for advanced AI | Voluntary (international) | Organisations developing advanced AI systems |
For the wider picture across jurisdictions, see our AI regulation by country comparison, and our detailed guides on the AI Promotion Act and AI governance in Japan by industry.
Two practical points follow. First, the enforceable obligations come from existing law applied to AI, not from the AI Promotion Act, which creates no sanctions: the personal-data law (APPI), the Copyright Act and its Article 30-4 training exception, and sector rules. Second, the soft-law layer carries real weight in Japan. The AI Guidelines for Business are treated as the standard of care, courts and regulators look to them, and non-cooperation with government measures can lead to being named publicly rather than fined.
In short, Japan rewards responsible adoption and leans on reputation rather than penalties. For an organisation, that makes a documented, defensible governance framework the practical way to meet expectations. Our AI GRC guide covers how those pieces fit together, and a short governance assessment benchmarks where an organisation stands against a structured model.
Yes. The AI Promotion Act (Act No. 53 of 2025) is Japan's first national AI law, passed on 28 May 2025 and promulgated on 4 June 2025. But it is a promotion and framework law: it sets basic principles and creates governance machinery, such as the AI Strategy Headquarters and an AI Basic Plan, rather than imposing enforceable duties on businesses.
No. The Act imposes no fines, no prohibitions and no mandatory conformity assessment. Its only business-facing duty is a non-binding obligation to endeavour to cooperate with government AI measures (Article 7). Enforcement is limited to information-gathering, guidance, and potentially naming non-compliant parties publicly, rather than sanctions.
The rules that can be enforced come from existing law applied to AI: the Act on the Protection of Personal Information (APPI) for personal data, the Copyright Act (including the Article 30-4 exception for AI training), and sector rules. The AI-specific layer, meaning the AI Promotion Act, the AI Guidelines for Business and the Basic Plan, is principles and soft law.
The AI Guidelines for Business are Japan's central voluntary framework, issued jointly by METI and MIC and now at Version 1.2 (March 2026). They consolidate earlier development, use and governance guidelines and function as the de facto standard of care for responsible AI, but they do not carry legal force on their own.
Generally yes. Article 30-4 of the Copyright Act permits using copyrighted works for information analysis, including AI training, without the rightsholder's authorisation where the purpose is not to enjoy the work's expression. It is subject to a proviso that the use must not unreasonably prejudice the interests of the copyright owner, which is the focus of ongoing debate and guidance.
The Financial Services Agency has taken a principles-based, non-binding approach through a discussion paper on the sound use of AI in finance, first published in March 2025 and updated to version 1.1 in March 2026, rather than issuing new binding rules.
A short, free assessment benchmarks where your organisation stands against a structured AI governance model, useful precisely because so much of Japan's AI layer is voluntary standard of care rather than hard law.
This page is general information describing the state of Japan AI policy as at 22 July 2026, not legal or compliance advice. Policy and regulatory positions change; always confirm the current position against the primary sources linked above and obtain advice from your own qualified counsel before relying on it.