含羞草社区

Japan became the second major Asia-Pacific economy to enact AI-specific legislation last summer with the Japan Act on Promotion of Research, Development and Utilisation of Artificial Intelligence-Related Technologies (Act No.53 of 2025).

Japan’s sober strategy should be a welcome approach to companies developing, providing and using AI, given sensational reversals in the EU AI Act’s framework, patchwork US State AI laws, and the penalties, registration requirements and broad compliance regimes of other developed economies.

In comparison, Japan has deliberately constructed a layered governance architecture that combines its new promotional statute, detailed voluntary guidelines, and the reliable enforcement power of existing law.

The result is a system thoughtfully designed, light enough to attract and support investment and research activity; flexible, with clear foundational bodies and principles to be elevated into place as international norms and domestic risks crystallise; and predictable, allowing for substantial adaptive lead time and reserving hard legal consequences only for areas of well-established and confirmed law.

Three layers, one strategy

The framework statute. The potential impact of AI on developed economies requires the highest level of government attention. But the actual shape of that impact is far from a settled outcome.

The definition of AI-related technology in the AI Promotion Act reflects this broad impact: technology that substitutes for human cognitive reasoning and judgement abilities, as well as information processing systems that use such technology.

Recognising this dynamic, the Act establishes the necessary top-level architecture without prematurely prescribing rules, creating an AI Strategic Headquarters within the Cabinet Office, chaired by the prime minister.

It first met in September 2025, approving an AI Basic Plan in December, anchored by a five-year, JPY1 trillion (USD6.3 billion) public investment commitment beginning in fiscal year 2026.

However, this framework carries no impulsive or looming sanctions, instead imposing only a “co-operation duty” on AI business users to align with emerging government policy.

Elevation layer. Where the Act sets the direction, voluntary guidelines do the gradual lifting, for example:

1. 1. Article 13 of the Act provides for the government to establish guidelines ensuring appropriateness of AI research, development and use. Under that power, the AI Strategic Headquarters immediately published the Guidelines for Ensuring Appropriateness (December 2025).
   2. The Digital Agency’s government procurement guidelines (May 2025) establish a chief AI officer in every ministry and a procurement check sheet that evaluates AI suppliers on governance, data handling, output quality and risk management.
   3. The AI Guidelines for Business version 1.1 (METI and MIC, March 2025) are currently the most detailed articulation of government expectations, organising guidance around three actor categories – AI developers, providers and business users – while applying a risk-based approach that scales governance to each system’s capabilities and context.

A separate AI Contract Checklist (METI, February 2025) provides clause-by-clause guidance for structuring AI development and service agreements across three procurement models.

These instruments are not legally binding. But they are not optional either. Over time, as courts become accustomed to the new technologies, compliance with guidelines may increasingly be treated as evidence of reasonable care under Japan’s general tort framework (Civil Code, Articles 709 and 715), meaning that organisations not implementing the voluntary layer can still adjust their posture before facing exposure under any future hard-law layer.

Enforcement backstop. In the areas where guardrails are well tested, Japan’s existing statutes apply to AI systems with full force. As prominent examples: the Act on the Protection of Personal Information (APPI) governs data collection, purpose limitation, cross-border transfers, and prompt inputs; the Copyright Act governs training data and output infringement; the Product Liability Act imposes strict liability for AI embedded in products; the Unfair Competition Prevention Act protects trade secrets; and the Antimonopoly Act addresses competition concerns.

Hard edges

The perception that Japan is “soft on AI” masks a few sharp enforcement edges. Four areas deserve particular attention.

Data protection is tightening. The APPI applies where personal information or personal data is involved across every stage of the AI lifecycle – training data collection, model development, inference, prompt inputs and outputs.

On 9 January 2026, the Personal Information Protection Commission published its Policy Direction to amend the APPI, proposing administrative monetary penalties for the first time. Japan currently has no general administrative monetary penalty regime under APPI. This change would alter the enforcement calculus.

The same Policy Direction proposes additional targeted consent exemptions for certain data use cases centred on statistical information creation (including AI development), new protection for children’s data, and classification of certain biometric-related data as sensitive. If enacted, implementation can be expected around 2027.

Separately, the government’s earlier caution regarding DeepSeek – alerting that its data is stored on servers in China, subject to Chinese law – signals that vendor location is a concrete compliance risk factor.

Copyright’s safe harbour is narrowing. Copyright Act Article 30-4 provides one of the world’s broadest exceptions for AI training, permitting use of copyrighted works for information analysis without rights-holder consent. But the exception is being tested.

In August 2025, three of Japan’s largest newspaper groups – Yomiuri Shimbun, Nikkei, and Asahi Shimbun – filed separate lawsuits against Perplexity AI alleging unauthorised “scraping” of more than 100,000 articles. Separately, the government asked that OpenAI address copyright concerns on its Sora video-generation model.

The Agency for Cultural Affairs has, meanwhile, published its General Understanding on AI and Copyright in Japan (May 2024), clarifying that while training may be permissible, reproducing protected creative expression can infringe copyright.

The Article 30-4 proviso is likely where the next round of litigation will be fought. This generally untested proviso bars the use of copyrighted works that would “unreasonably prejudice” rights holders’ interests.

Competition enforcement is arriving. The Japan Fair Trade Commission published a Generative AI Report in June 2025 analysing market concentration, data access barriers and the potential for algorithmic collusion. It also moved to investigate AI search services and their use of news content. Companies building or deploying AI systems that process competitor data or influence market pricing should expect increasing scrutiny.

Sector regulators are engaging. The Financial Services Agency published an AI discussion paper in March 2025 (updated in March 2026) setting expectations for model risk management and explainability in financial institutions. Healthcare AI is subject to the Pharmaceutical and Medical Device Act’s pre-market review for AI-enabled medical devices. The trend is clear: general AI guidelines are being translated into sector-specific expectations.

What to do now

Six priorities stand out for businesses in Japan’s AI market. First, map roles under the AI Guidelines – developer, provider or business user – recognising that many organisations hold multiple roles simultaneously, and implement the governance expectations for each.

Second, build an AI inventory with internal risk classifications. Japan mandates no prescribed risk taxonomy, so these must be developed independently.

Third, structure vendor contracts using the METI’s contract checklist, paying particular attention to input data handling, output ownership, restrictions on use of input data for model training, and audit rights.

Fourth, implement trade secret controls for AI tool use. Inputting confidential information into external AI services risks destroying the “secrecy management” status required for protection under the Unfair Competition Prevention Act.

Fifth, if selling to government, prepare for the procurement check sheet and establish relationships with ministry chief AI officers.

Sixth, for AI research laboratories and companies working with them, engage with the AI Safety Institute’s evaluation frameworks and monitor the draft Principle Code for advanced AI systems, which is now in public consultation.

Looking ahead

Several developments will shape the near-term landscape.

The APPI amendment bill, if enacted, will introduce Japan’s first administrative monetary penalty regime for data protection violations and create more tailored consent exemptions for certain AI-related data uses. Organisations should be prepared for this simultaneous tightening and loosening.

The Principle Code for advanced AI may also impose more concrete expectations on foundation model developers and providers.

Tokyo District Court proceedings against Perplexity, meanwhile, will begin to define the practical boundaries of Article 30-4’s copyright training exception.

The Economic Security Promotion Act’s designation of “specified critical technologies” may additionally extend to advanced AI, with implications for export controls and international research collaboration.

Japan’s AI governance is often misread as incomplete. It is, more accurately, deliberately adaptive: designed to promote innovation while preserving the capacity to regulate forcefully when needed.

The question for businesses is not whether Japan’s approach will harden, but only at what pace, and in which areas. Companies building governance programmes around the voluntary framework will find hardening of Japan’s guardrails, when it comes, to be manageable and expected. Companies mistaking “soft” for “optional” will not.

Kojima Law Offices
Akasaka Tameike Tower 7F
Akasaka 2-17-7 Minato-ku
Tokyo 107-0052 Japan
Tel: +81 3 3222 1401
Email: godsoe@kojimalaw.jp