Japan's AI Promotion Act entered into full force on 1 September 2025, establishing a framework-based, soft-law regime that sets expectations for responsible AI development and use without imposing criminal or administrative penalties at this stage. The first AI Basic Plan, adopted in December 2025, translates those framework principles into concrete policy priorities and signals the direction of future, harder regulation. Foreign businesses operating in Japan should treat this period as a structured window to build compliant governance practices before more prescriptive rules follow.
Understanding the Legal Architecture
Japan has deliberately chosen a principle-led, non-punitive approach to AI regulation – at least in its opening phase. The AI Promotion Act, in force since 1 September 2025, does not replicate the risk-tiered, penalty-backed model of the EU AI Act. Instead, it articulates broad duties of care around trustworthy AI, aligns with the Hiroshima AI Process principles, and creates a coordination mandate across ministries. Compliance, for now, is measured against guidelines and codes of practice rather than hard legal thresholds.
The AI Basic Plan – Japan's first – was adopted in December 2025 and performs a different function. It sets government-wide investment priorities, identifies sectors where AI deployment will be actively promoted (including healthcare, infrastructure and manufacturing), and provides a roadmap that private actors can use to anticipate where sector-specific rules are likely to emerge. For foreign businesses, the Basic Plan is as important as the Act itself: it reveals where regulatory attention will concentrate next.
If your business involves AI law matters in Japan, understanding the interaction between the framework Act and the Basic Plan is the essential starting point. Our team advises on both instruments as a connected whole.
Who Is Affected: Scope and Applicability
The AI Promotion Act applies broadly to AI developers, deployers and providers of AI services operating within Japan, including foreign companies whose AI systems are made available to users in Japan. There is no explicit extraterritorial provision equivalent to the EU AI Act's market-reach test, but the practical effect is similar: if your product or service touches Japanese users or Japanese data, the Act's principles apply to how you design, document and deploy that system.
Key categories of actors caught by the framework include:
- Developers of foundation or general-purpose AI models – including large language models offered via API or integrated into commercial products sold into the Japanese market.
- Deployers of high-risk AI systems – entities embedding AI into consequential decision-making in areas such as credit assessment, employment screening or medical support tools.
- AI service providers operating platforms – businesses that host or distribute third-party AI applications to Japanese consumers or business clients.
- Foreign companies with Japanese subsidiaries or branches – the Act looks to the point of deployment and user impact, not only to the legal seat of the developer.
Employment-related AI applications deserve specific attention. Algorithmic recruitment tools, performance monitoring systems and automated scheduling carry distinct obligations under existing Japanese labour law, which intersects with the AI framework. Our employment law practice in Japan advises on this intersection regularly.
Step-by-Step Compliance Procedure
Step 1 – Map Your AI Systems Against the Framework
Begin with a structured inventory of all AI systems you develop or deploy in Japan. For each system, document: the function it performs, the data it processes, the decisions it influences, and the category of users it affects. The AI Promotion Act does not prescribe a formal mapping format, but voluntary guidelines issued by the Ministry of Economy, Trade and Industry (METI) provide a useful template. This inventory becomes the foundation for every subsequent compliance step.
Step 2 – Assess Alignment with Basic Plan Priorities
Cross-reference your AI systems against the sector priorities identified in the December 2025 AI Basic Plan. Systems deployed in promoted sectors may attract both greater regulatory scrutiny and greater governmental support. Where your system falls into a promoted area, consider engaging early with the relevant ministry or industry association – Japan's soft-law model rewards proactive dialogue in ways that a pure penalty-based regime does not.
Step 3 – Establish an Algorithmic Transparency Framework
Algorithmic transparency is a recurring theme in both the Act and the Basic Plan. This does not yet mean mandatory algorithmic audits by a public body, but it does mean maintaining documentation sufficient to explain, at minimum to affected users and, where challenged, to courts or regulators, how key decisions are reached. For general-purpose AI models, this includes records of training data provenance, known limitations and intended use cases.
Step 4 – Review Data Processing Under Existing Law
The AI Promotion Act operates alongside, and does not displace, Japan's Act on the Protection of Personal Information (APPI), the Unfair Competition Prevention Act and sector-specific rules administered by regulators such as the Financial Services Agency (FSA Japan). A compliance programme that addresses only the AI Act without auditing personal data flows, third-party data licensing and biometric processing will be incomplete. The FSA Japan has separately signalled expectations around AI use in financial services that go beyond the general framework.
Step 5 – Implement Internal Governance Structures
Japan's framework model places significant weight on corporate self-governance. At a minimum, foreign businesses should establish: a designated AI governance lead or committee; an internal review process for new AI deployments; a mechanism for receiving and acting on user complaints about AI-driven decisions; and a periodic review cycle tied to updates in the Basic Plan or ministerial guidelines. The absence of such structures is increasingly noted as a governance deficiency in transactional due diligence.
Step 6 – Prepare for Dispute Resolution
AI-related disputes in Japan currently proceed through ordinary civil litigation – the Tokyo District Court handles the majority of complex commercial disputes – or through arbitration, with the Japan Commercial Arbitration Association (JCAA) an established forum for international cases. There is no dedicated AI tribunal. Documenting your compliance steps carefully creates a contemporaneous record that is material if a dispute arises over liability for AI-generated harm.
If you would like to discuss your current AI governance posture in Japan, please contact our team at info@ferrazwhitmore.com or visit our contacts page.
Documents and Records to Maintain
Because the regime is currently soft-law, there is no prescribed filing with a centralised AI Office equivalent. However, the following documentation is recommended as both a compliance safeguard and a litigation-preparedness measure:
- AI system inventory with risk classification and intended use description
- Records of training data sources, licensing arrangements and data quality assessments
- Human oversight protocols for high-risk AI systems, including escalation procedures
- User-facing disclosure materials explaining AI involvement in decisions
- Incident logs recording any AI malfunction, bias event or user complaint
- Records of periodic internal review and any corrective action taken
- Correspondence with METI, FSA Japan or other regulators, including voluntary disclosures
Realistic Timelines
The AI Promotion Act has been in force since 1 September 2025. The AI Basic Plan, adopted December 2025, will be reviewed and updated on a cycle that is expected to follow government budget cycles. Sector-specific implementing regulations – where they emerge – are likely to be consulted on publicly before enactment, providing a window of several months between draft and force. Foreign businesses should therefore not wait for hard rules to begin building governance frameworks; the expectation embedded in Japan's soft-law approach is that responsible actors will be compliant before compulsion arrives.
For businesses entering Japan now, a realistic compliance build-out – covering inventory, documentation, governance structure and staff training – typically runs over several months depending on the complexity of the AI systems involved. Businesses with existing Japan operations should treat the period from September 2025 through 2026 as a maturation phase rather than a grace period.
Cost Drivers
Because the current regime imposes no penalty-based compliance costs, the principal cost drivers are internal and advisory rather than regulatory. These include: legal counsel to map AI systems against the framework and existing Japanese law; technical advisory for documentation and model card preparation; staff training on internal governance procedures; and, where relevant, costs associated with adapting user interfaces to meet transparency expectations. Companies with complex general-purpose AI deployments or significant data processing operations will face proportionately higher documentation and review costs.
Common Errors by Foreign Clients
Several patterns recur when foreign businesses first engage with Japan's AI framework:
- Treating soft law as no law. The absence of penalties does not mean the absence of obligation. Courts, regulators and business partners will assess your governance against the framework principles, and gaps create both legal and reputational exposure.
- Ignoring the intersection with APPI and sector rules. AI compliance in Japan cannot be siloed from personal data compliance. Systems that process biometric or sensitive data are subject to existing APPI requirements irrespective of what the AI Act does or does not say.
- Assuming EU AI Act compliance is sufficient. Japan's framework reflects different principles – it is less prescriptive on conformity assessment but places greater weight on corporate self-governance and voluntary engagement with regulators. EU-certified systems may still require additional documentation and disclosure work for the Japanese market.
- Overlooking employment law dimensions. AI tools used in hiring, performance management or workforce scheduling engage Japanese labour law principles independently of the AI framework. A number of disputes in this area have already reached mediation or court proceedings.
- Underestimating the value of early regulatory engagement. Japan's ministry-led soft-law model responds well to proactive outreach. Companies that engage early with METI or sector regulators typically find it easier to shape expectations around their specific use case.
Comparative Context: Regional Regulatory Developments
Japan's approach contrasts with regulatory developments elsewhere in the region. For reference, on 7 September 2026 the Supreme People's Court of China issued judicial opinions on adjudicating AI-related disputes – 24 provisions in five parts – applying existing law including the Civil Code, Cybersecurity Law, Copyright Law and Personal Information Protection Law. Those opinions addressed: personality rights infringement arising from unauthorised replicas of faces and voices; fault-based liability with notice-and-action duties for AI service providers; algorithmic price discrimination; and AI-generated false information. This was judicial guidance, not new legislation, but it signals how courts in the region are extending existing legal frameworks to cover AI-specific harms.
For businesses managing multi-jurisdictional AI deployments across Asia-Pacific, understanding how different frameworks interact is essential. Our analysis of AI dispute adjudication developments in China provides further comparative context.
Decision Checklist for Foreign Businesses
- Have you inventoried all AI systems deployed or available to users in Japan?
- Have you assessed each system against the principles in the AI Promotion Act and the sector priorities in the AI Basic Plan?
- Have you reviewed data processing activities under APPI and applicable sector rules (including FSA Japan guidance where relevant)?
- Have you established an internal AI governance structure with a designated lead?
- Do you maintain documentation sufficient to explain AI-driven decisions to users, courts or regulators?
- Have you assessed whether employment-related AI tools engage Japanese labour law obligations?
- Have you considered proactive engagement with METI or relevant sector regulators?
- Have you reviewed your dispute resolution readiness, including whether your contracts designate JCAA arbitration or Tokyo District Court jurisdiction for AI-related disputes?
This checklist is a starting point rather than a definitive compliance assessment. The specific obligations applicable to your business will depend on the nature of your AI systems, your sector and your corporate structure in Japan. To discuss your situation in detail, please write to info@ferrazwhitmore.com or reach us through our contacts page.
How Ferraz & Whitmore Can Assist
Our Japan practice advises foreign companies at each stage of the compliance process described above: from initial system mapping and gap analysis, through documentation frameworks and governance design, to representation in regulatory dialogues and, where disputes arise, support in proceedings before the Tokyo District Court or in JCAA arbitration. We work alongside local Japanese counsel and technical advisers to provide coordinated guidance that covers both the legal and operational dimensions of AI compliance in Japan. Enquiries are welcome at info@ferrazwhitmore.com.
Disclaimer: This publication is provided for informational purposes only and does not constitute legal advice. The information herein should not be relied upon as a substitute for professional legal counsel tailored to your specific circumstances. Ferraz & Whitmore assumes no liability for actions taken or not taken based on the contents of this material. For advice regarding your particular situation, please contact info@ferrazwhitmore.com.