The Japan AI Promotion Act — formally the Act on Promotion of Research, Development and Utilisation of AI-Related Technologies — is Japan's first comprehensive AI statute. The National Diet passed it on 28 May 2025; its principal provisions entered into force on 4 June 2025, with full effect reported in November 2025. This page summarizes the Act as compiled from secondary analyses (Future of Privacy Forum, CSIS, Bird & Bird, and others); the primary text was led through the policy process by the Cabinet Office.
- Passed: 28 May 2025 (National Diet)
- Effective: 4 June 2025 (principal provisions); full effect November 2025
- Enacting body: National Diet; Cabinet Office led policy; AI Strategic Headquarters established by the Act
Summary of the framework
The Act takes an innovation-first approach, with no risk tiers, no specific business obligations beyond "endeavouring to cooperate" with government measures, and no penalties. Governance is centralised in the AI Strategic Headquarters within the Cabinet, chaired by the Prime Minister. Commentators describe the statute as structurally and philosophically opposed to the EU AI Act (EU AI Act (Regulation 2024/1689)) and the South Korea AI Basic Act (South Korea AI Basic Act — Source Summary), which impose risk-tier classifications and binding obligations.
The Act's structure rests on a coordinating body and an operational plan rather than on enforceable duties:
- The AI Strategic Headquarters (within the Cabinet), chaired by the Prime Minister with all Cabinet ministers as members, serves as the coordinating body and formulates and implements a national AI Basic Plan.
- The AI Basic Plan is the operational document developed by the Headquarters.
- Duties on businesses are framed as soft law ("endeavour to cooperate").
- The Act establishes no risk-tier framework, no individual rights, and no penalties.
Several existing instruments operate alongside the statute. METI's AI Guidelines for Business (v1.1, with an update expected in 2026) function as the operational soft-law baseline. The Japan AI Safety Institute (AISI), housed under the IPA since 2024, continues with an expanded remit and conducts safety evaluations. Government procurement leverage and reputational incentives serve as de facto enforcement in the absence of statutory penalties. The Act includes a five-year review clause.
Obligations are stated as endeavours rather than mandates. Under Article 14, businesses bear a 努力義務 ("endeavour to cooperate") with government measures, with no specific audits, impact assessments, or compliance duties. Under Article 15, citizens are to understand AI's nature and limitations and engage responsibly. On safety, the government is to issue non-binding guidelines, and AISI Japan is to conduct safety evaluations.
Key claims
- The Diet passed the AI Promotion Act on 28 May 2025; principal provisions entered into force 4 June 2025 (high confidence).
- The Act establishes the AI Strategic Headquarters within the Cabinet, chaired by the Prime Minister with all Cabinet ministers as members (high confidence).
- The Act contains no penalties — financial or otherwise — for non-compliance; enforcement rests on cooperation, reputation, and government-procurement signals (high confidence).
- Business obligations are framed as an "endeavour to cooperate", a Japanese soft-law formulation (high confidence).
- The Act creates no individual rights, including no right to explanation, contest, or human review of AI-mediated decisions (high confidence).
- The first AI Basic Plan was expected in 2026 but had not been finalised as of the source date (medium confidence).
- Japanese officials and commentators contrast the approach with the EU AI Act; the CSIS analysis calls it "agile governance" that "fosters a global nexus through pluralistic interoperability" (high confidence).
Relation to other AI-governance instruments
The statute presents a soft-law model that several analyses position against the binding regimes of the EU AI Act (EU AI Act (Regulation 2024/1689)) and the South Korea AI Basic Act (South Korea AI Basic Act — Source Summary), as well as alongside other national frameworks such as Brazil AI Bill (PL 2338/2023) — Source Summary, Canada AIDA (Bill C-27, Part 3) — Source Summary, and Australia Voluntary AI Safety Standard — Source Summary. Its pro-innovation framing parallels that of America's AI Action Plan, though the AI Action Plan is executive action rather than statute, and that of Singapore's model framework (Singapore Model AI Governance Framework for Generative AI (2024)).
Japan led the drafting of the international G7 Hiroshima Code of Conduct for Advanced AI (2023), which is substantively stricter than the domestic statute; one is an international voluntary instrument and the other a domestic statute, both non-binding in different ways. The Act also connects to the soft-law approach described in AI Safety Cases and Frameworks and to Japanese positioning on AI Sovereignty. It serves as the legislation page for the Act (Japan AI Promotion Act — Source Summary).
Provenance
Secondary-source compilation; no PII, credentials, or sensitive identifiers appear in the source material.
Sources
- Government of Japan, "Act on Promotion of R&D and Utilization of AI-Related Technology Now in Full Effect" (Highlighting Japan, Nov 2025):
https://www.gov-online.go.jp/hlj/en/november_2025/november_2025-08.html - Future of Privacy Forum, "Understanding Japan's AI Promotion Act: An 'Innovation-First' Blueprint for AI Regulation"
- CSIS, "Japan's Agile AI Governance in Action: Fostering a Global Nexus Through Pluralistic Interoperability"
- Bird & Bird, "Japan's AI Governance — Major Government Steps Since the AI Act" (2025)
- International Bar Association, "Japan's Emerging Framework for Responsible AI"
- IAPP, "Global AI Governance Law and Policy: Japan"
- Nemko Digital, "How Japan is regulating AI: Inside the AI Promotion Act"
- Regulations.ai listing for the Act:
https://regulations.ai/regulations/RAI-JP-NA-PRDUAXX-2025