Japan’s 2025 AI Promotion Act: Structuring Innovation Through Soft Regulation

1. A Purposeful and Strategic Legislative Framework

On 15 March 2025, the Japanese National Diet enacted the Act on the Promotion of Research, Development and Utilization of Artificial Intelligence-Related Technologies, commonly referred to as the AI Promotion Act. Promulgated on 25 March 2025, the law is scheduled to enter into force on 1 July 2025.

This landmark legislation was introduced by the Cabinet Office with the ambition to provide a national framework to support AI development, not by imposing restrictive obligations, but by establishing principles of coordination, government support, and ethical guidance. The law reflects Japan’s long-standing emphasis on fostering technological innovation while preserving flexibility for industry actors.

The objectives of the Act are articulated through a set of guiding principles, including:

• The promotion of research and innovation in AI-related technologies;
• Respect for human dignity and social trust in AI applications;
• Encouragement of public-private cooperation and open innovation;
• Harmonization with international norms and cooperation with foreign partners;
• Development of human capital and digital literacy across the population.

The text of the Act is composed of 17 articles, structured around four key axes: philosophical foundations, stakeholder responsibilities, government support, and institutional governance.

‍

2. Responsibilities and Institutional Architecture

While the law imposes no binding obligations on private sector actors, it invites all stakeholders—businesses, academic institutions, citizens—to “cooperate” with government-led AI strategies. This cooperation is intentionally left undefined and is not enforceable as a legal duty. Instead, the Act relies on soft law techniques and voluntary alignment with government-issued guidance.

The responsibilities are distributed as follows:

• National Government:
  • Defines and implements Japan’s AI strategy;
  • Issues guidelines and supports international coordination;
  • Collects data on harmful uses and investigates social impacts of AI;
  • Funds research and infrastructure programs.
• Local Governments:
  • Promote regional use of AI in public services and innovation clusters.
• Private Entities and Academia:
  • Expected to support national AI initiatives through voluntary cooperation;
  • Encouraged to contribute to ethical and transparent use of AI.
• Citizens:
  • Invited to strengthen their AI literacy and awareness of algorithmic impacts.

The law establishes a National AI Strategy Council, comprising representatives of government, academia, industry, and civil society. This body will:

• Monitor implementation of national policy;
• Advise the Prime Minister on updates to the basic strategy;
• Provide a forum for stakeholder consultation and forward-looking evaluation.

The Prime Minister is empowered to issue and revise a Basic AI Promotion Policy, which sets the tone and operational priorities of the national approach.

‍

3. Government Support and Remedial Mechanisms

The AI Promotion Act introduces a series of public support measures, aimed at reinforcing Japan’s position in the global AI landscape. These include:

• Research Funding: Establishment of programs to finance basic and applied AI research;
• Talent Development: Creation of training programs to cultivate AI-related skills across disciplines;
• Infrastructure Access: Provision of shared access to computing power, large datasets, and testing environments;
• Ethical Safeguards: Monitoring of AI-related incidents where individual rights or public interests are compromised.

Where AI causes harm or poses risks to fundamental rights, the government is mandated to:

• Collect relevant information;
• Conduct contextual analysis;
• Implement remedial measures where necessary.

However, no enforceable rights or sanctions mechanisms are provided for individuals adversely affected by AI systems. Unlike the EU’s AI Act, the Japanese law does not introduce transparency, explainability, or redress obligations for high-risk AI deployments. This absence of binding constraints has been a point of concern for several digital rights groups and legal scholars in Japan.

‍

4. Complementarity with Article 30-4 of the Copyright Act

Although the 2025 AI Promotion Act does not specifically address the use of data for AI training, it must be read in parallel with Article 30-4 of the Japanese Copyright Act, introduced in 2018.

This article allows the use of copyrighted works for “information analysis”, including AI training, without prior authorization, provided:

• The use is not intended for the enjoyment of the expressive content;
• It does not unreasonably prejudice the legitimate interests of the copyright owner.

In practice, this has allowed both commercial and non-commercial developers to train AI models on large corpora of copyrighted material, even where the source may not be lawfully licensed—so long as the purpose is analytical and not consumptive. The lack of enforceable restrictions on dataset provenance has contributed to Japan’s image as a jurisdiction particularly favorable to AI development.

However, the 2024 administrative guidance issued by the Japanese Agency for Cultural Affairs emphasized that:

• The exception does not apply if technological protection measures are circumvented;
• The use of pirated or infringing content may result in liability if done knowingly.

This context reinforces the Japanese model as one of broad permissiveness, coupled with general principles of good faith and minimal oversight.

‍

5. A Legal Model Distinct from the European Union

Compared to the European Union’s AI Act, which creates detailed compliance obligations, mandatory risk assessments, and potential fines for violations, Japan’s 2025 approach is not prescriptive but promotional. It seeks to cultivate a domestic AI ecosystem by:

• Reducing legal uncertainty for developers;
• Avoiding over-regulation at an early stage of market maturity;
• Encouraging self-regulation and public trust rather than imposing statutory duties.

This model may attract investment and research activity but remains exposed to criticism for its lack of enforceable rights and accountability mechanisms.

‍

This legal commentary was prepared by a French attorney specialized in intellectual property, artificial intelligence, and emerging technologies. It is not intended to constitute legal advice on Japanese law. Any legal analysis of the AI Promotion Act should be sought from a qualified Japanese practitioner.