AI Governance in Japan by Industry: Finance, Healthcare, Manufacturing, and the Soft Law Approach

Japan's AI governance is characterised by voluntary principles, sector-led guidance, and a deliberately non-prescriptive regulatory approach. This industry guide covers what Japanese companies and international companies operating in Japan actually need to do.

Key Takeaways

Japan's AI governance approach is 'soft law' — voluntary guidelines, industry self-regulation, and ministry guidance rather than binding legislation. This creates more flexibility but less certainty than the EU approach.
The AI Promotion Act (May 2025) establishes basic principles for AI governance but is explicitly non-prescriptive — it creates a framework for voluntary compliance rather than mandatory requirements.
The Financial Services Agency (FSA) has issued specific guidance on AI in financial services — risk management, explainability, and third-party AI vendor oversight are the primary expectations.
The Ministry of Economy, Trade and Industry (METI) AI governance guidelines for business are the primary cross-sector reference for Japanese companies — they align closely with OECD AI Principles.
The Personal Information Protection Commission (PPC) has issued guidance on AI and the Act on the Protection of Personal Information (APPI) — the third amendment to APPI in 2022 created specific obligations relevant to AI.

"Apenas para fins informativos. Este artigo não constitui aconselhamento jurídico, regulatório, financeiro ou profissional. Consulte um especialista qualificado para orientação específica."

Japan's AI governance framework — what industry needs to know in 2026

Japan's AI governance landscape changed materially in 2025. The country moved from a purely soft-law approach to having its first AI-specific statute: the AI Promotion Act (Act on Promotion of Research and Development and Utilization of AI-Related Technologies), passed by the Diet on 28 May 2025. Most provisions entered into force on 4 June 2025, with the AI Strategic Headquarters and AI Basic Plan provisions following on 1 September 2025. The first AI Basic Plan was approved by Cabinet on 23 December 2025. Japan is now the third major jurisdiction — after the EU and South Korea — with AI-specific legislation, though its approach is deliberately different from both.

Japan's stated ambition is to become "the world's most AI-friendly country" — a phrase used repeatedly across official documents. The approach balances innovation promotion with governance through guidelines rather than prescriptive regulation.

The AI Promotion Act — what it does and doesn't do

The AI Promotion Act is a "fundamental law" — legislation that establishes principles, national objectives, and institutional architecture rather than prescriptive rules with penalties. It adopts a multi-stakeholder model, assigning roles to central and local governments, academia, businesses, and citizens. It does not impose direct fines or penalties on businesses. It does not ban any AI applications. It does not require conformity assessments or pre-launch registration.

What it does: defines AI as technologies simulating human cognition; positions AI as a driver of public welfare and economic resilience; establishes the AI Strategic Headquarters under the Cabinet Office (first meeting 12 September 2025); mandates the creation of an AI Basic Plan; requires the government to take legislative and financial actions concerning AI; and empowers the Headquarters to request cooperation from any entity when deemed necessary.

The AI Utilization Guidelines, published by the Headquarters on 19 December 2025 pursuant to Article 13, provide more detailed operational guidance.

The AI Guidelines for Business — the operational reference

The METI/MIC AI Guidelines for Business (Version 1.1, 28 March 2025) are the primary operational compliance reference. While non-binding, they function as the standard of care that courts, regulators, and commercial counterparts reference. They follow a three-tier structure: foundational values (human dignity, inclusion, sustainability); ten cross-sector principles (fairness, privacy, safety, security, transparency, accountability, literacy, fair competition, innovation); and practical tools including checklists and case studies.

The Guidelines tailor expectations to three roles: developers (who build AI), providers (who offer AI services), and business users (who deploy AI in their operations). A key feature is the call for executive-level responsibility — AI governance should be embedded into organisational governance like cybersecurity.

For companies operating in Japan, the Guidelines operate on a "comply or explain" basis. Non-compliance doesn't trigger fines, but courts and regulators increasingly treat adherence as the standard of care. In January 2026, Chief Cabinet Secretary Minoru Kihara announced the first use of investigative powers under the Act: a review of AI-generated sexual deepfake content. The enforcement model is reputational rather than financial — "name and shame" rather than fines.

AI Basic Plan (December 2025)

The AI Basic Plan approved 23 December 2025 sets medium-term goals around four pillars: adoption of AI in government services and across the economy; domestic AI development through investment in research, data, and computing infrastructure; trustworthy AI through safety, security, and ethics measures; and international AI governance leadership through the Hiroshima AI Process.

The fiscal year 2026 budget reflects these priorities with significant allocations to AI infrastructure, government AI utilisation, and governance implementation. The Cabinet Office is leading a ¥22 billion investment in generative AI for medical diagnostic support. More than ¥10 trillion in public support is planned over seven years through FY2030 for AI-semiconductor infrastructure.

Sector-specific considerations

Financial services. The Financial Instruments and Exchange Act requires algorithmic high-speed trading registration, risk management systems, and transaction records. Financial Services Agency guidance applies to AI in banking, insurance, and securities.

Healthcare. AI medical devices are regulated under the Pharmaceuticals and Medical Devices Act (PMD Act). The ¥22 billion generative AI investment for medical diagnostics signals sectoral priority.

Government procurement. The Digital Agency's May 2025 guideline on procuring and using generative AI inside government requires each ministry to appoint a Chief AI Officer (CAIO). Procurement includes governance checkpoints, high-risk review logic, and explicit consideration of data sovereignty (servers outside Japan).

Copyright. The Cultural Affairs Agency published "General Understanding on AI and Copyright in Japan" (May 2024). Japan's copyright framework has been relatively permissive for AI training, but the 2025 Yomiuri Shimbun lawsuit against Perplexity signals that rights holders are testing boundaries.

How Japan compares to EU and other jurisdictions

The EU AI Act is prescriptive with risk classification, conformity assessments, and substantial penalties. Japan's AI Promotion Act is a framework law with guidelines and existing sectoral law for operational effect. A company deploying AI across both must comply with both — Japan's framework does not satisfy EU AI Act obligations and vice versa.

South Korea's AI Basic Act (effective 22 January 2026) takes a middle path — more prescriptive than Japan with mandatory risk assessment and transparency for high-impact AI, but less prescriptive than the EU. Singapore, like Japan, emphasises voluntary frameworks but with more sector-specific binding requirements (MAS TRM Guidelines for financial services).

What companies operating in Japan should do

Align internal AI governance with the METI/MIC Guidelines (Version 1.1). Develop risk documentation covering bias, safety, reliability testing. If selling to Japanese government, review the Digital Agency's generative AI procurement guidance. Include AI-specific clauses in supplier and vendor contracts. Monitor the AI Strategic Headquarters' output — comprehensive guidelines are expected on a rolling basis through 2026. For companies also operating in the EU, build a unified governance framework satisfying both Japan's guidelines and EU AI Act requirements.

Primary sources: METI — Ministry of Economy, Trade and Industry · White & Case — Japan AI Tracker · IAPP — Japan AI Governance