Laws/Regulations directly regulating AI (“AI Regulations”)

Japan currently has no law specifically directed to regulating AI. Thus, at this time, Japan is taking an indirect approach to support the policy goal of prioritizing innovation while minimizing foreseeable harms.

To this end, the Ministry of Economy, Trade and Industry (METI) published Japan's AI regulatory policy (AI Governance Report 1.1) (the METI Guidelines) in January 2022.¹ The METI Guidelines are not legally binding but are expected to support and induce voluntary efforts in good practices by companies developing and deploying AI systems. Separately, Japan has also published the Hiroshima International Guiding Principles for Organizations Developing Advanced AI Systems (the Hiroshima Principles), which aim to establish and promote guidelines worldwide for safe, secure and trustworthy AI.²

A working group of the governing Liberal Democratic Party is reportedly advancing a proposal, titled "the Basic Act on the Advancement of Responsible AI," that would adopt a "hard law" approach to regulate certain generative AI foundation models (Proposed AI Bill). Under the proposed law, the government would designate the AI system and developers subject to regulation, impose obligations on them with respect to vetting, operation and output of the systems, and require periodic reports. The proposed obligations would likely be similar to the "voluntary commitments" that some major AI companies in the US (such as OpenAI, etc.) made to the White House in July 2023. It is thought that the law would provide a general framework, while industry groups and/or individual developers would be expected to establish the specific standards by which they would comply. It is further thought that the government would have the authority to monitor AI developers and impose fines and penalties for violations of the reporting obligations and/or compliance with the substance of the law. Thus, if the Proposed AI Bill is passed on these terms, this would mark a shift from a "soft law" approach to a "hard law" approach to regulating AI in Japan.

Status of the AI Regulations

As noted above, at this time, there are no laws or regulations in Japan that expressly govern the development, use, or provision of AI. Note, however, the Proposed AI Bill (see above) may result in a "basic law" directed at AI-based systems and applications.

Other laws affecting AI

There are various laws that, though not adopted specifically to regulate AI, are likely to affect the development or use of AI in Japan. A non-exhaustive list of key examples includes:

• The Digital Platform Transparency Act, which imposes requirements on large online malls, app stores, and digital advertising businesses to ensure transparency and fairness in transactions with business users
• The Financial Instruments and Exchange Act, which requires businesses engaging in algorithmic high-speed trading to register with the government, establish a risk management system, and maintain transaction records
• The Civil Code, which allows tort claims to be brought against a person who has given instructions to AI to produce defamatory content about another person and published such content
• The Copyright Act and the Act on the Protection of Personal Information, which can be applied to inappropriate uses of AI

In addition, various acts involving AI may already be caught by the Criminal Code: Using AI to defame another could fall under the ambit of defamation; using AI-generated fake content to interfere with someone's business could be punishable as an obstruction of business; and giving unauthorized commands to another person's computer can also be punishable under the criminal code.

Definition of “AI”

There is no legally recognized definition of "AI" in Japan. The Draft AI Guidelines for Business published by METI and the Ministry of Internal Affairs and Communications ("Draft AI Guidelines for Business") note that there is currently no established definition for AI and that it is difficult to strictly define its scope. However, for the purpose of the Draft AI Guidelines for Business document, METI states that AI is an abstract concept that includes AI systems themselves, as well as machine learning software and programs.³

Territorial scope

As noted above, at this time, there are no laws or regulations in Japan that expressly govern the development, use, or provision of AI. The non-binding METI Guidelines do not refer to a territorial scope or contemplate extra-territorial jurisdiction; nor do they distinguish between on-shore and off-shore entities or actions. However, the proponents of the Proposed AI Bill are considering the appropriate territorial scope for a regulatory system.

Sectoral scope

As noted above, at this time, there are no laws or regulations in Japan that expressly govern the development, use, or provision of AI. Accordingly, there is no specific sectoral scope at this stage.

The Draft AI Guidelines for Business are broadly applicable to AI developers, AI providers and those using AI in the private sector.

Compliance roles

As noted above, at this time, there are no laws or regulations in Japan that expressly govern the development, use, or provision of AI. Accordingly, there are currently no specific or unique compliance obligations imposed on developers, users, operators and/or deployers of AI systems.

Core issues that the AI Regulations seek to address

As noted above, at this time, there are no laws or regulations in Japan that expressly govern the development, use, or provision of AI. Nevertheless, the Hiroshima Principles identify a number of significant risks, including disinformation, copyright, cybersecurity, risks to health and safety, and societal risks such as the ways in which advanced AI systems can give rise to harmful bias and discrimination.

Further regarding copyright, in-depth discussions are being held in Japan about how existing laws (i.e., the Copyright Act of Japan) should address issues concerning rights and harms that may arise from Generative AI. Also, the Council of the Agency for Cultural Affairs is expected to announce its position regarding copyright where AI is trained on the works of humans and AI-generated content.

Risk categorization

As noted above, at this time, there are no laws or regulations in Japan that expressly govern the development, use, or provision of AI. In addition, AI or AI systems are not generally classified according to risk in the relevant guidelines and principles announced in Japan. However, it has been reported that some government officials believe that a risk-based approach is necessary.⁵

Key compliance requirements

As noted above, at this time, there are no laws or regulations in Japan that expressly govern the development, use, or provision of AI.

Nevertheless, the METI Guidelines note that certain AI principles adopted by Japan are moving toward a consensus. Those principles are:⁶

• Human-centric – i.e., the utilization of AI must not infringe upon the fundamental human rights guaranteed by the constitution and international standards
• Education/literacy – i.e., policy makers and managers of businesses involved in AI must have an accurate understanding of AI, as well as suitable knowledge and ethics to permit the appropriate use of AI in society
• Privacy protection – i.e., each stakeholder must consider whether to handle personal data with greater discretion, in accordance with the level of importance and sensitivity of the data
• Ensuring security – i.e., society must be aware of the balance between the benefits and risks of AI, and endeavor to improve social safety and sustainability as a whole
• Fair competition – i.e., a fair competitive environment between countries and stakeholders regarding the use of AI must be maintained
• Fairness, accountability and transparency – i.e., in an "AI-Ready Society," it is necessary to ensure fairness and transparency in decision-making, appropriate accountability for the results, and trust in the technology, so that people who use AI are not subject to undue discrimination with regard to personal background, or to unfair treatment in terms of human dignity
• Innovation – i.e., equal collaboration among universities, research institutions and companies regarding AI research should be promoted, and governments should proceed with regulatory reforms to reduce impeding factors and establish an efficient and beneficial society with the aid of AI technologies

Regulators

As noted above, at this time, there are no specific laws or regulations in Japan that directly regulate AI, and so there is no specific AI regulator. Nevertheless, the following ministries and agencies are substantively engaged in establishing and promoting guidelines regarding AI:

• The Ministry of Economy, Trade and Industry
• The Ministry of Internal Affairs and Communications
• The Agency for Cultural Affairs (particularly for copyright issues)
• The Personal Information Commission

Guidelines promulgated by ministries in Japan are often followed closely by companies and the public, even though they are not binding law.

Enforcement powers and penalties

As noted above, at this time, there are no laws or regulations in Japan that expressly govern the development, use, or provision of AI. As such, enforcement and penalties relating to the creation, dissemination and/or use of AI are governed by related violations in non-AI legislation (such as those set out in Item 3 above).⁷ However, the Proposed AI Bill contemplates implementing a regulatory system that includes fines and penalties for violations.

1 See the METI Guidelines here.
2 See the Hiroshima Process International Code of Conduct for Organizations Developing Advancing AI Systems here.
3 See the "Draft AI Guidelines for Business" (here), page 8.
4 See the "Draft AI Guidelines for Business" (here), page 3 and the Appendices here.
5 See here.
6 See the METI Guidelines (here), page 9: "As mentioned above, the discussion on what AI Principles are is moving toward a general consensus. Japan adopts seven principles in "Social Principles of Human-centric AI" (decision of the Integrated Innovation Strategy Promotion Council)", and the Social Principles of Human-Centric AI here.
7 In addition to enforcement set out in item 3: METI is authorized to issue a recommendation or an order to a large digital platform provider if they do not disclose certain factors used to determine search rankings; the Personal Information Commission (the PIC) is authorized to supervise and monitor businesses handling personal or other related information, and it may issue a recommendation or an order; the PIC may publish the name of businesses that failed to comply with its order; and, while the Criminal Code does not specifically focus on AI, as discussed above, activity using AI may constitute a crime and be punished.

Maoko Kamiya (Associate, White & Case, Tokyo) and Jeames Keate (Associate, White & Case, Tokyo) contributed to this publication.