Este artigo está disponível apenas em inglês no momento.
Japan's AI Promotion Act 2025: The World's Most Innovation-Friendly AI Law
Japan passed its first dedicated AI law in May 2025 — the AI Promotion Act. It has no penalties, no prohibitions, and no mandatory conformity assessments. But METI guidelines carry real weight, and the new AI Strategic Headquarters chaired by the Prime Minister signals Japan's serious approach to AI governance.
Key Takeaways
Japan's AI Promotion Act (May 28, 2025, full enforcement September 2025) is the world's first major national AI law with no penalties, no mandatory compliance obligations, and no prohibited AI uses.
The Act establishes the AI Strategic Headquarters chaired directly by the Prime Minister — the highest-level AI governance body of any major economy.
METI and MIC's AI Guidelines for Business (Version 1.1, March 2025) are the operative compliance benchmark. Courts and regulators treat them as the standard of care for AI deployment in Japan.
Non-compliance with guidelines triggers 'name and shame' enforcement — public naming of non-compliant operators — rather than financial penalties. Reputational risk is real.
Japan's copyright exception under Article 30-4 of the Copyright Act permits AI training on copyrighted materials subject to the 'unreasonably prejudice' limitation — creating opportunities but also emerging litigation risk.
"Apenas para fins informativos. Este artigo não constitui aconselhamento jurídico, regulatório, financeiro ou profissional. Consulte um especialista qualificado para orientação específica."
Japan's AI Promotion Act: what it is and what it is not
On May 28, 2025, Japan's Parliament enacted the Act on Promotion of Research, Development and Utilization of Artificial Intelligence-Related Technologies — commonly called the AI Promotion Act. It entered full enforcement on September 1, 2025. This is Japan's first dedicated AI legislation, and it deliberately differs from the EU's AI Act in almost every significant respect.
The AI Promotion Act does not prohibit any AI applications. It does not require conformity assessments, technical documentation, or pre-deployment registration. It carries no financial penalties for non-compliance. It is, as described by one analysis, "the world's most innovation-friendly AI law from a major economy."
What it does is establish the legal and institutional infrastructure for Japan's AI governance: the AI Strategic Headquarters (chaired by the Prime Minister and including all Cabinet ministers), the authority to issue an AI Basic Plan, and the framework for sector-specific guidance and oversight. Article 25(2) gives the Headquarters the ability to request cooperation from any entity and to publicly disclose the names of non-cooperating operators — the "name and shame" enforcement mechanism.
The AI Strategic Headquarters: governance at the highest level
Japan's AI governance structure is unusually centralised at the top. The AI Strategic Headquarters is chaired directly by the Prime Minister and includes all Cabinet ministers. This structure ensures that AI governance in Japan is not fragmented across agencies — it is a national priority issue tied to economic security, digital infrastructure, and social stability.
The Headquarters published the AI Utilization Guidelines on December 19, 2025, pursuant to Article 13 of the Act. The AI Basic Plan, ratified for implementation through 2026, outlines Japan's strategic objectives including accelerated AI adoption across manufacturing, healthcare, transportation, and small business sectors to address demographic challenges.
METI guidelines: the operative compliance standard
While the AI Promotion Act creates the statutory framework, the practical compliance benchmark for companies operating in Japan is the METI/MIC AI Guidelines for Business, Version 1.1, published March 28, 2025. These guidelines define the standard of care that courts, regulators, procurement officers, and enterprise buyers apply when assessing AI governance.
The guidelines are structured around 10 cross-sector principles — fairness, privacy, safety, security, transparency, accountability, and others — and apply differentially to developers, providers, and users of AI systems. They call for executive-level responsibility: AI governance should be embedded like cybersecurity, into organisational structure rather than treated as a compliance checklist.
A "comply-or-explain" standard applies in practice: companies interacting with Japanese government, major enterprise customers, and international partners are expected to demonstrate alignment with the METI guidelines or explain deviations. Non-alignment is not illegal but it is commercially significant.
Japan's copyright training data exception
One feature that made Japan attractive to AI developers globally is Article 30-4 of Japan's Copyright Act, which permits AI training on copyrighted materials. The exception is broader than equivalent provisions in other jurisdictions, which contributed to Japan's reputation as AI-training-friendly.
The exception is not unlimited. The "unreasonably prejudice" proviso applies when training use unreasonably harms the copyright holder's interests. This limitation is now attracting litigation scrutiny — the Yomiuri vs. Perplexity case is being watched as a potential test of where the exception ends. Companies relying on Article 30-4 for large-scale AI training should monitor Japanese court filings in 2026.
Sector-specific obligations
Japan's healthcare AI is regulated by the Pharmaceuticals and Medical Devices Agency (PMDA). AI used in medical devices requires approval, and AI supporting clinical decisions must comply with professional responsibility standards that keep final accountability with licensed clinicians. Financial AI is supervised by the FSA with expectations for internal governance frameworks covering stress testing and bias monitoring. The AI Promotion Act explicitly preserves and builds on this sector-specific regulatory architecture rather than replacing it.