Japan has no EU-style AI law with fines. Its main statute, the AI Promotion Act (Act No. 53 of 2025), sets goals, creates an AI Strategic Headquarters led by the Prime Minister and asks businesses to cooperate with government measures. The detail sits in a national plan and in voluntary guidelines. The Cabinet adopted the second edition of the AI Basic Plan on 14 July 2026. The Act text we read has no penalty provisions.
01How does Japan approach AI rules?
Japan leans on promotion, planning and voluntary guidance more than on bans. In the Cabinet Office's English draft of the AI Basic Plan, the government says it combines measures under the AI Act with guidelines to nudge businesses and citizens toward steps they take on their own initiative (draft plan).
This page is general information, not legal advice. A company working in Japan should ask a Japanese lawyer.
02What does the AI Promotion Act say?
The Act is Act No. 53 of 2025, promulgated on 4 June 2025, and its Japanese title translates roughly as the Act on the Promotion of Research and Development and Utilization of AI-Related Technologies. The official text on e-Gov is in Japanese. It does the following:
- Article 1 says its purpose is to set basic principles, a national basic plan and a headquarters for promoting AI.
- Article 7 says businesses that use AI must cooperate with national and local government measures. Article 8 asks citizens to deepen their understanding of AI.
- Article 16 tells the state to collect information and to study cases where AI used for an improper purpose or in an improper way harmed people's rights, then give guidance, advice and information to developers and business users.
- Articles 19 to 28 set up the AI Strategic Headquarters, headed by the Prime Minister (Article 22).
The parts of the Act on the basic plan and the headquarters came into force later by Cabinet order, within three months of promulgation. The text we read contains no fines or other penalties.
03What is the AI Basic Plan, and what is new in 2026?
The Cabinet decided the first AI Basic Plan on 23 December 2025 and the second edition on 14 July 2026, according to the Cabinet Office's AI strategy page. The page's newest item, dated 18 September 2026, announces the AI-related budget requests for the next fiscal year.
We could not read the final English text of the second edition. The Cabinet Office's English draft from 19 June 2026 is a provisional translation, so treat its details as a guide. It says AI risks have become more complex with agentic AI, which acts toward goals on its own. It lists technical risks such as chains of failures, social risks such as unclear responsibility and effects on jobs, and national security risks such as cyberattacks that AI carries out by itself (draft plan).
The draft frames its goal as Trustworthy AI and says 180,000 government employees already use the government AI platform called GENAI. It promises to strengthen Japan's AI Safety Institute, which evaluates AI models technically, to prepare guidelines on rights infringements linked to AI, and to keep reviewing the AI Act itself as risks change.
04How do public bodies and other countries fit in?
The draft plan asks the government to write rules for how administrative bodies introduce AI agents, and to set out points to note when using AI in public administration so that citizens' rights, transparency and fairness are protected (draft plan).
It also stresses working abroad. Japan led the Hiroshima AI Process, an international effort on how AI should be governed, and the draft says Japan will expand that work. The AI Safety Institute takes part in an international network that exchanges views with its counterparts in the United States and the United Kingdom on evaluating AI and on standards. These are policy aims in a draft, not rules that bind companies.
05Which guidelines do companies follow?
Two sets of guidance matter. The AI Strategic Headquarters decided its guidelines on ensuring appropriateness in AI research, development and use on 19 December 2025, as the draft plan records. The Ministry of Economy, Trade and Industry and the Ministry of Internal Affairs and Communications publish the AI Guidelines for Business, and METI's guidelines page lists version 1.1 as its newest text.
Given the government's stated approach of encouraging voluntary steps, read these as expectations and good practice, not as statutes. We did not read the guidelines in full.
?What is not known
- The final English text of the second AI Basic Plan, and any changes from the June 2026 draft.
- Whether the government will add binding rules. The draft plan says it will keep reviewing systems, including the AI Act, but gives no decision.
- Whether METI has issued a newer version of the business guidelines than the one its page lists.
QQuestions people ask
Does Japan fine companies for misusing AI under the Act?
The Act text we read has no penalty provisions. The plan notes public concern about portrait-right and copyright issues, which fall under other laws that we did not review.
Is Japan's AI Safety Institute a regulator?
The draft plan describes it as the body responsible for technical evaluation of AI models, and says it will be strengthened. It does not describe it as a regulator.
Last checked October 24, 2026. Rules and settings change, so open the source to see the current version.
- Act on the Promotion of Research and Development and Utilization of AI-Related Technologies (Act No. 53 of 2025) e-Gov Law Search, Government of Japan The official Japanese statute; used for the articles described.
- AI Strategy, Cabinet Office Cabinet Office, Government of Japan The government's AI page with topics and dates, including the 14 July 2026 decision.
- Artificial Intelligence Basic Plan (draft, provisional English translation) Cabinet Office, Government of Japan English draft dated 19 June 2026; used for plan themes.
- AI Guidelines for Business Ministry of Economy, Trade and Industry Lists the business guidelines; used for version 1.1.
General information, not legal, medical, financial or investment advice. We are not affiliated with any company, agency or standards body named here. See our editorial policy and corrections, or tell us about a mistake at [email protected].