Explanation of Japan's AI Bill: Aiming for a Balance Between Innovation and Safety

This content was created using generative AI tools. As there is a possibility that the detailed content may contain inaccuracies or misinformation, please refer to other sources when making important decisions.

Introduction

On February 28, 2025, the Japanese government approved the 'Bill on the Promotion of Research, Development, and Utilization of Artificial Intelligence (AI) Related Technologies' (commonly known as the AI Bill), which is Japan's first basic law on AI, and submitted it to the ordinary session of the Diet. This bill establishes a fundamental framework to address both the opportunities and risks arising from the rapid development of AI technology. We will explain the features and impact of this bill, which aims to make Japan 'the world's most AI-friendly country for development and utilization.'

Summary of this article

1. Background: Triggered by the emergence of risks accompanying the rapid development of generative AI, Japan's lag in AI investment and utilization, and growing public demand for AI regulation.

2. Features:

   • A basic law (promotion law) without penalties, balancing the promotion of innovation with ensuring safety.

   • Establishment of an 'Artificial Intelligence Strategy Headquarters' (headed by the Prime Minister) within the Cabinet to formulate a 'Basic Plan' that is reviewed annually.

   • Respecting voluntary initiatives rather than imposing excessive regulations on companies.

3. Main Measures:

   • Promotion of research and development support, infrastructure development, and human resource development.

   • Active participation in the formation of international rules.

   • Monitoring of AI risks and investigation/analysis of serious incidents.

   • Guidance and advice to businesses as necessary.

4. Comparison with other countries:

   • EU: The 'AI Act' imposes strict regulations based on risk

   • USA: A trend toward deregulation at the federal level, with individual states enacting their own laws

   • Japan: Positioned between the two, seeking a unique cooperative model

5. Challenges:

   • Ensuring the effectiveness of laws without penalties

   • Consideration for small and medium-sized enterprises and startups

   • Concerns about over-regulation and excessive surveillance

Background: Why an AI bill now?

The environment surrounding AI is changing rapidly. Since the emergence of ChatGPT at the end of 2022, the capabilities of generative AI have improved dramatically, bringing major transformations to industry and daily life, while risks such as fraudulent websites, scams using synthetic voices, and the spread of misinformation have also become apparent.

Movements toward AI regulation are progressing in countries around the world, with various approaches being taken, such as the enactment of strict AI regulations (AI Act) in the EU (European Union) and the issuance of AI ethics guidelines and executive orders in the United States.

On the other hand, when compared internationally, Japan's lag in the AI field is noticeable. In 2023, private AI investment in Japan was approximately $700 million (12th in the world), compared to approximately $67.2 billion in the US and approximately $7.8 billion in China. Furthermore, the usage rate of generative AI is 9% for individuals and 47% for companies, which is significantly lower than in the US (46% for individuals, 85% for companies) and China (56%, 84%).

According to a KPMG survey, the percentage of people who answered that 'AI needs regulation' was highest in Japan at 77%, and the percentage who answered that 'I think AI can be used safely under current rules and laws' was the lowest at 13%, indicating that public anxiety about AI is also rising.

Against the backdrop of these social demands, the government recognized the need for a new legal system that balances the promotion of innovation with risk management, leading to the formulation of the AI bill.

Overview and Features of the AI Bill

1. Scope and Basic Principles

The bill covers 'technology for realizing functions that substitute for human cognitive, reasoning, and judgment capabilities,' in other words, AI algorithms and systems in general that substitute for human intellectual work.

The following points are set forth as basic principles:

• AI technology is the foundation for economic and social development and is also important for national security

• Maintaining and strengthening research and development capabilities and improving industrial international competitiveness

• Comprehensive and planned promotion from basic research to practical application

• Ensuring transparency in the research, development, and utilization of AI

• Proactive participation in the formation of international AI rules

2. Promotion Structure (Governance)

An 'Artificial Intelligence Strategy Headquarters' will be established within the Cabinet as a command center to comprehensively promote AI policy. It is a powerful structure in which the Prime Minister serves as the head, the Chief Cabinet Secretary and the Minister in charge of Artificial Intelligence Strategy (newly established) serve as deputy heads, and all cabinet ministers participate as members.

The Strategy Headquarters will formulate and publish an 'Artificial Intelligence Basic Plan' and promote various measures in line with it. This plan is scheduled to be reviewed annually in accordance with technological progress, enabling flexible responses.

3. Content of Basic Measures

The bill stipulates the following as measures to be taken by the government:

• Promotion of R&D and utilization of results: Financial support, project promotion, and social implementation of results

• Infrastructure development and human resource development: Development of intellectual infrastructure such as facilities, equipment, and data, and promotion of education

• Participation in international initiatives: Active participation in international rule-making and reflection in domestic guidelines

• Risk monitoring and consideration of countermeasures: Information gathering on domestic and international AI trends, and investigation and analysis of major incidents

• Response to businesses and the public: Guidance, advice, and information provision as necessary

4. Responsibilities of Stakeholders

This bill defines the responsibilities of various entities:

• National and local governments: Policy promotion in accordance with basic principles

• R&D institutions: Ensuring safety and returning results to society

• AI-utilizing businesses: Active use of AI and cooperation with government measures

• The public: Deepening understanding and interest in AI and cooperating with measures

5. Legal Binding Force and Degree of Regulation

This law has a strong character as a "promotion law" and does not contain direct penalty provisions. While the government can exercise investigative authority and provide guidance or advice for correction to businesses, there is no mechanism for penalties if businesses do not comply.

In malicious cases, measures such as publicizing company names are envisioned, but the basic stance is to prioritize industrial development and respect voluntary initiatives as much as possible.

Additionally, the supplementary provisions of the law include a review clause, which explicitly states that measures such as legal amendments will be taken in response to changes in international trends and social conditions.

Impact on Society and Industry

Impact on Corporate Activities

This bill does not impose new licensing procedures or direct regulatory costs on companies, but rather focuses on building a foundation to promote AI utilization. Compared to strict regulations in Europe, the short-term increase in burden is considered small.

Government support for AI research and development, infrastructure development, and human resource cultivation will serve as a tailwind for private companies, and the promotion of innovation is expected.

On the other hand, companies are also required to strengthen AI governance. There is an obligation to cooperate with national policies, and it will be necessary to develop risk management systems and respond to information provision requests in accordance with guidelines formulated by the government.

Impact on Technology Development

The measures in this bill are expected to promote domestic AI technology development. Research and development support, strengthening industry-academia-government collaboration, and human resource development measures will enrich the soil for the birth of new AI technologies and services.

By having the government indicate priority areas and budget allocations in the AI field every year based on a basic plan, it will become easier for the private sector to make investment decisions with a view to the future. If a framework for risk management is established, social acceptance will increase, and companies will be able to conduct technology demonstrations and deploy new services with peace of mind.

Impact on Labor, Education, and Healthcare

The full-scale utilization of AI is expected to bring about changes in various fields of society:

• Labor field: While productivity improvements and the resolution of labor shortages are expected, there are also concerns about job displacement in some occupations. The government's human resource development measures aim to boost reskilling and the development of highly skilled personnel, thereby increasing the ability to respond to employment shifts.

• Education field: While the use of individual learning support AI and teacher burden reduction tools is progressing, issues such as dependence on AI and misinformation are also emerging. The government will play a role in guiding the use of AI in a direction that brings out human creativity through the presentation of guidelines and training support.

• Healthcare field: As diagnostic imaging AI, drug discovery AI, and virtual nurses become more widespread, the government will investigate and analyze any troubles that may occur and consider necessary measures. An overarching supervisory system will lead to the building of trust among patients and medical professionals, enabling safe and sustainable medical AI utilization.

Issues and Concerns

Questions about Effectiveness

This bill is a basic law without penalties, and there are points raised such as, 'If there are no penalties even if businesses do not cooperate, can it be called sufficient regulation?' Establishing mechanisms to obtain cooperation for investigations (for example, through industry associations) and strengthening the system to increase the effectiveness of the law are future challenges.

Consideration for Over-regulation

It has also been pointed out that there is a risk of leading to excessive regulation depending on future implementation. If the government demands detailed technical standards or prior notification on the grounds of ensuring safety, it could become an excessive burden for small and medium-sized enterprises and startups, potentially hindering innovation.

The government is required to steer carefully so as not to mistake the 'balance between strengthening regulation and fostering industry.' The philosophy of this bill also states that 'legal system responses should be limited to cases where voluntary efforts by businesses cannot be expected,' and it is important to adhere to this principle.

Impact on Small and Medium-sized Enterprises and Startups

Consideration for small and medium-sized enterprises and startups with limited resources is also essential. For companies with limited personnel and funds necessary for AI governance, it may be difficult to develop sophisticated systems.

It is hoped that the government will enhance support measures for small and medium-sized enterprises (such as establishing consultation counters for governance and providing toolkits) and strive to realize the purpose of this law fairly.

Concerns about Increased Surveillance

There are concerns that the government gaining the authority to grasp the actual state of AI development and usage, as well as investigative powers, could lead to excessive surveillance. Balancing this with the protection of corporate intellectual property and user privacy is a challenge.

The government is required to demonstrate transparency regarding the scope and procedures for exercising its investigative powers and to establish mechanisms to prevent the creation of an arbitrary surveillance society.

Comparison with other countries

EU: Strict risk-oriented comprehensive regulation

The EU's 'AI Act' is a strict framework that classifies AI systems into four categories based on the level of risk, imposing different regulatory measures for each:

• Unacceptable risk: Prohibited from use, such as social scoring and indiscriminate facial recognition systems

• High risk: Strict requirements imposed on AI use in critical infrastructure, recruitment, etc.

• Limited risk: Transparency obligations imposed on chatbots, etc.

• Minimal risk: No special regulations for everyday AI

In the event of a violation, massive fines of up to 6% of global turnover are being considered, leading to concerns about potential isolation (Galapagos syndrome).

USA: Respect for voluntary principles and localized regulation

The US does not have a comprehensive AI law at the federal level. While guidelines such as the 'AI Bill of Rights' and 'Executive Order on AI' were presented under the Biden administration, the shift to the Trump administration in 2025 has steered the direction toward deregulation and the encouragement of development.

At the state level, unique legal frameworks are being developed, such as Utah's 'Artificial Intelligence Policy Act' (requiring disclosure that AI is AI) and Colorado's 'Artificial Intelligence Act' (preventing discrimination by high-risk AI).

Japan: An intermediate approach and a unique balance

Japan's AI bill is positioned between the EU and the US in terms of regulatory intensity. It does not establish strict pre-market regulations like the EU, nor does it leave everything to the market like the US model, but is instead exploring a unique model of 'government governance support + voluntary initiatives'.

A distinctive feature is the creation of an agile structure, such as the establishment of a strategic headquarters with all cabinet ministers participating and a basic plan that is reviewed annually. The clear goal of 'aiming to be the country where AI is easiest to use in the world' is also a sign of the commitment to improving the environment and promoting investment.

Future outlook: International cooperation and Japan's role

AI is a global technology, and international cooperation is essential. Japan is leading international governance discussions through the 'Hiroshima AI Process' and the GPAI (Global Partnership on Artificial Intelligence).

Japan's AI bill has an open-platform character, with the flexibility to incorporate or relax regulatory elements as needed. By leveraging this, it is possible to update domestic systems in accordance with international consensus.

If Japan can build a 'global model that balances risk response and innovation promotion,' international cooperation will further advance, contributing to a sustainable future where the benefits of AI can be enjoyed safely.

Summary

Japan's AI bill establishes a basic framework that balances the promotion of AI innovation with the assurance of safety. By taking a middle ground between the strict regulations of Europe and the self-regulation of the United States, Japan aims for its own unique approach of 'flexible and effective AI governance.'

The bill is merely a starting line, and it is important to continue implementing and improving specific measures from here on out. The challenge moving forward will be whether we can safely and effectively advance social transformation through AI, based on a cooperative relationship between the government and the private sector.

A system is being put in place for the public and private sectors to work together to promote AI strategy, with the goal of becoming 'the country in the world where AI is easiest to develop and utilize.' We hope that this bill will serve as a catalyst to accelerate AI utilization in Japan, realizing both economic growth and the resolution of social issues.

#AIBill #ActOnPromotingResearchDevelopmentAndUtilizationOfArtificialIntelligenceRelatedTechnologies #GenerativeAI #AIStrategyCouncil #AIBasicPlan #ChatGPT #AIGovernance #InnovationPromotion #RiskResponse #AIRegulation #EU #AIAct #USA #HiroshimaAIProcess #GPAI #AIBillOfRights #AIExecutiveOrder #SoftLaw #SMEs #Startups #Reskilling #HumanResourceDevelopment #ResearchAndDevelopment #InternationalCooperation #AIEthics #TransparencyEnsurance #InvestigativeAuthority #OpenPlatform #WorldModel #CountryWhereAIIsEasiestToUtilize