Japanese AI Regulation and Competition Law

The Network Law Review is pleased to present a special issue entitled “The Law & Technology & Economics of AI.” This issue brings together multiple disciplines around a central question: What kind of governance does AI demand? A workshop with all the contributors took place on May 22–23, 2025, in Hong Kong, hosted by Adrian Kuenzler (HKU Law School), Thibault Schrepel (Vrije Universiteit Amsterdam), and Volker Stocker (Weizenbaum Institute). They also serve as the editors.

**

Abstract

This article explores Japan’s emerging approach to regulating generative AI, balancing innovation and risk. The government emphasizes soft, sector-specific regulation guided by international norms, while the 2025 AI Bill outlines policies promoting AI development and safety. The Japan Fair Trade Commission (JFTC) is conducting a market study to assess competitive concerns in generative AI, and its new Mobile Software Competition Act (MSCA) provides an ex-ante framework to curb anti-competitive practices in mobile software. Together, these initiatives highlight Japan’s commitment to fostering AI growth while maintaining fair competition and international regulatory alignment.

*

1. Introduction

While the potential benefits to society and economies of generative AI are enormous, it takes no Aristotle to understand the technology’s risks that have prompted various governments to consider or introduce new regulations. Japan is no exception. While the Japanese government understands the importance of not hindering innovation, like many of its counterparts, it is carefully deliberating how to create a regulatory system that ensures confidence for the technology’s expansion. Policymakers have great expectations for AI’s effects on the economy, and competition policies play a crucial role in harnessing and proliferating its full potential. This article provides a brief overview of current deliberations and decisions on AI regulation in Japan, focusing on the relationship with competition law, especially the Japan Fair Trade Commission (JFTC)’s ongoing market study on generative AI (a type of artificial intelligence that can create new content). It also provides an overview of Japan’s first “ex-ante” digital competition framework — the Mobile Software Competition Act (MSCA) — which was enacted in June 2024 and is expected to be fully enforced by December 18, 2025, and what touchpoints it has to AI-related anticompetitive conduct.

2. The Japanese Government’s Approach to AI Regulation

Advancements in AI now enable handling vast amounts of information across nearly all fields of human activity. In the near future, AI will undoubtedly make significant contributions to society by greatly improving efficiency in various industries and convenience in everyday life. Its diverse risks are becoming increasingly apparent, however. These risks include AI-generation for fraudulent websites, confusion due to widespread misinformation, and national security threats, such as the development of CBRNs (weapons using Chemical, Biological, Radiological, and Nuclear materials) and unprecedented cybersecurity challenges.

In light of such circumstances, the Japanese Government established the “AI Institutional Study Group” as the driving research body under the AI Strategic Council within the Cabinet Office to help determine high level approaches for artificial intelligence regulation in Japan.[1] The AI Institutional Study Group has held hearings with various stakeholders including business operators, independent experts, and local governments as part of its efforts to determine what legal updates are required for Japan to thrive in the AI era, including if an entirely new framework is necessary.[2] The AI Strategic Council and the AI Institutional Study Group jointly published an Interim Report on February 4, 2025 that highlights the importance of balancing the promotion of innovation and risk mitigation (the “Interim Report”).[3] The following are taken directly from the English version summary of the Interim Report:

• Balancing innovation promotion and risk mitigation
  • Promoting innovation including support to research and development, human resources development, and expansion of data and computing resources
    • According to the Interim Report, the Japanese government is already working on organizing and providing high-quality Japanese language data to companies, promoting the establishment of data centers, securing power supplies, and supporting AI semiconductors, as well as implementing human resource development projects to support young researchers and doctoral students in the field of next-generation AI. The Interim Report advocates for such assistance for research and development of AI to continue.[4]
  • Appropriate combination with laws and guidelines
    • The Interim Report offers a sector-specific approach. In order to avoid stifling innovation, the government is to primarily rely on existing legal frameworks such as the Civil Code, the Penal Code, the Act on the Protection of Personal Information, and the Copyright Act, and respond to new challenges through soft law such as guidelines.[5] For example, in its effort to mitigate legal concerns, in March 2024, the Cultural Affairs Agency articulated its interpretation of the relationship between AI and the current Copyright Act for potential copyright infringements when generative AI models absorb vast amounts of data to generate content.[6]
  • Use of the common principles such as the OECD AI principle and Hiroshima Process International Guiding Principles for Organizations Developing Advanced AI Systems and individual existing laws
• International cooperation
  • Leading discussion toward formulation of AI governance
    • The Interim Report highlights the achievements of the Hiroshima AI Process under Japan’s presidency, and proposes that the Japanese government make efforts to take a leading role in discussions at various international conferences based on the concept of the Hiroshima AI Process.[7]
  • Ensuring international consistency and interoperability
    • The Interim Report emphasizes the importance of ensuring international consistency and interoperability so that the requirements of international norms related to safety and other matters align with those applied in Japan. This would enable business operators in Japan to expand smoothly into overseas markets and allow its citizens to access AI services globally.[8]

The publication signals a softer regulatory stance than what Japan previously signaled^[9] in 2024. Policymakers have recently advocated for prompt legalization of concepts above to avert hard regulations on AI, and instead strengthen the government’s function to formulate softer high level strategies to improve safety and foster growth. A bill on AI rules based on the proposals from the Interim Report above received cabinet approval, leading to a potential new law outlined in section II below.

3. Japanese Bill on AI

The Japanese Cabinet approved a bill on February 28, 2025 that strives to promote both the development and utilization of AI while ensuring safety (the “AI Bill”). As of mid-May of 2025, the AI Bill is under deliberation in the national diet.[10] The outline of the AI Bill is as follows:

• Establish basic principles for the development and use of AI
• In view of the importance of AI-related technology in economic society and national security, Japan will seek to maintain its R&D capabilities and enhance its international competitiveness.
• The government will promote AI research and development comprehensively and strategically, from basic research to practical application.
• The government will ensure transparency to facilitate appropriate R&D and utilization of AI.
• The government will make efforts to take a leading role in international cooperation.
• Establish an “AI Strategic Headquarters” headed by the Prime Minister and comprising all cabinet ministers, to take initiative in reducing costs and complexities of responsibly implementing AI and clarify interpretation of sector specific laws in areas relating to AI.
• Basic policies include:
  • Japan’s participation in the formulation of international AI norms
  • Development of domestic guidelines in line with international norms
  • The government shall collect information on domestic and international trends in the R&D and utilization of AI-related technologies, analyze cases in which the rights and interests of the citizens have been infringed, consider appropriate countermeasures based on such analysis, and conduct other necessary investigations and research. Based on these findings, the government shall provide guidance, advice, and information to business operators and citizens.
• Business operators have obligations to cooperate with measures taken by the government.

The AI Bill essentially sets out the government policy direction and aims to serve as a basic framework law. While it stipulates business operators’ obligations to cooperate with measures taken by the government, penalties for violation are not stipulated. This reflects the Japanese government’s position — consistent with the proposal in the Interim Report — that aims to avoid the stifling of innovation by primarily relying on existing laws and soft law measures. In this respect, Japan’s AI Bill seems to represent a different approach from the EU’s AI Act which regulates a wide range of AI systems and has hefty fines for violations.

In addition, given the recommendations in the Interim Report, the AI Bill can be seen as a law that provides a foundation for the development of government guidelines and for research to understand the actual use of AI. Therefore, the enactment of this law does not signify the completion of Japan’s AI regulatory framework; rather, it is middle way for continuous policy development through ongoing government research, application of existing sectoral laws, and cooperation with the private sector.

4. The JFTC’s Market Study

The Japanese AI Bill above provides a legal framework presenting basic policies that do not focus on competition policies. As Japan’s competition authority, the JFTC is publicly conducting a market study focused on generative AI.[11]^[12] As a first step, the Commission produced a discussion paper^[13] on October 2, 2024 and collected comments until late November of the same year. As market dynamics quickly change and levels of investment are unprecedented for this potentially society-altering technology, the JFTC took a step away from its traditional method of conducting market studies — which often takes more than a year to produce — and decided to nimbly research and produce its findings as the market evolves in an effort to continuously monitor the current situation and take measures when needed from a competition perspective.

The October 2024 discussion paper laid out the JFTC’s understanding of the current generative AI market structure after numerous discussions with stakeholders and experts. The following is how the JFTC categorizes generative AI markets for discussion purposes:

• Infrastructure Layer: computing resources (such as GPUs), data, expertise (specialized talent)
• Model layer: Large Language Models, other generative AI models
• Application Layer: integration with existing digital services
• Other important factors: cloud services,^[14] switching/migrating development environments, open-source/closed-source, and partnerships

The JFTC identified potential issues within this stack, and solicited comments related to Japanese competition policy:

• Access Restrictions and Exclusion of Competitors:

A handful of large enterprises are in advantageous positions for acquiring GPU and data necessary for the development of generative AI models. If access to these are restricted and competitors are excluded, new market entry opportunities and other competition dynamics may be impacted.

• Self-Preferencing

If a provider of generative AI models develops their model in such a way that their own products or services appear more favorably in the inference results compared to those of other companies, it could impact competition related to those products or services.

• Tying

If a dominant provider of a service bundles the use of its own generative AI model as a condition for offering that service, it could potentially impact competition related to generative AI models.

• Parallel Conduct Using Generative AI

While generative AI-driven price surveys can increase price competition, the alignment of underlying data and algorithms can lead to similar or identical pricing strategies and production targets by various companies, potentially impacting competition.

• Acquiring Specialized Talent Via Partnerships

If a company aims to corner highly skilled talent for themselves by forming partnerships, resulting in effects similar to a business transfer, it could impact competition.

As of mid-May of 2025, the JFTC plans on compiling its views after follow-up hearings, including with those who submitted comments, and releasing a paper with its findings. While the ongoing market study remains the predominant method of understanding and addressing generative AI issues, the JFTC has a powerful new framework to handle digital platforms on mobile devices, which many observe as the main avenue for dominant market players to roll out AI technologies.

5. The Mobile Software Competition Act (MSCA)

As the JFTC continues its market study on generative AI to ensure fair competition in a rapidly evolving industry, the role of digital platforms, particularly smartphones, in AI deployment becomes increasingly significant. The software environment of smartphones directly influences the distribution and utilization of AI models and applications, making it a crucial factor in competition policy.

On June 12, 2024, Japan’s Smartphone Software Competition Promotion Act, or Mobile Software Competition Act (MSCA) was passed in parliament, creating Japan’s first ex-ante competition framework. It will be fully enforced starting December 18, 2025 by the JFTC. The MSCA not only aims to foster competition in the mobile software market but also has the potential to support the healthy development of AI technology.

• Background:
  • The widespread use of smartphones has made them a cornerstone of daily life and economic activities in Japan, but specific software essential for smartphones (mobile operating systems, app stores, browsers, and search engines, collectively called “specified software” in the MSCA) is provided by a small number of dominant
  • Those dominant operators’ anticompetitive practices hinder fair and free competition in these markets. Yet, as entry of new competitors is unrealistic, and proving anticompetitive practices under the Antimonopoly Act is both resource intensive and protracted, restoring competition through traditional means is exceedingly arduous.
  • Creating a competitive environment for key smartphone software while ensuring security and safety is crucial for Japan. Breaking anticompetitive barriers to allow for innovation by a variety of organizations ultimately offers consumers a wide range of services.
• The MSCA in action:
  • The JFTC designates software service providers of each specified software whose businesses exceed a certain scale, as defined by a Cabinet Order.
    • The Order deemed business operators with specified software featuring 40 million average monthly domestic users[15] during the last fiscal year as “designated providers.”
    • These designated providers must adhere to certain prohibitions and obligations to address competition issues within their respective software for which they were designated.
  • The MSCA mandates designated providers to submit compliance reports, facilitates the sharing of information by stakeholders to the JFTC, and promotes cooperation among relevant ministries and agencies.
  • The law also grants the JFTC the authority to investigate, issue cease-and-desist orders, impose surcharge payment orders, and more.
  • The new regulatory framework will require continuous communication with stakeholders, including designated providers and application developers, to improve current business models.

Although the MSCA was drafted before the explosive proliferation of AI and did not explicitly take the technology’s existence in mind for its design, there are several touchpoints for the JFTC to catch anticompetitive conduct in the AI sphere.^[16]

Firstly, Article 7 Paragraph 2 of the MSCA regulates third-party apps’ access to mobile operating system functions. Designated providers on mobile operating systems will be obligated to provide certain interoperability to third parties, and apps that use AI assistant functions built into the operating system. In the future, as on-device AI processing becomes more common, it will become increasingly important to ensure interoperability to third party apps that perform AI processing on mobile operating system functions.

Article 12 Paragraph 1 Subparagraph (b) of the MSCA, meanwhile, stipulates that a choice screen must be displayed by the designated providers on mobile operating system for certain applications as a way to reduce the dominance of pre-installed apps and promote competition by offering users more control. The JFTC’s thinking is leaning towards AI assistant apps not falling within scope of choice screen obligations for now.

Another touchpoint may be Article 6 of the MSCA, which states that designated providers of mobile operating system and application stores under the law may not unjustly discriminate against, or unfairly treat application developers in transactions and usage conditions of mobile operating systems and application stores. Case-specific thinking on what is considered “unjust discrimination” or “unfair treatment” will be further clarified in guidelines to be published by JFTC, but as increasing amounts of AI technologies are provided through mobile apps, this may become a powerful tool in controlling anticompetitive practices to ensure further diversified proliferation of AI in Japan.

Article 7’s Paragraph 1 prevents designated providers from stopping third party providers from offering their own application stores. One can optimistically hope that this opens doors to innovation on the types of applications smartphones in Japan can leverage, including the newest AI boosted tools that previously may not have been developed or rolled out due to high intermediary fees or lack of appropriate audiences.

Finally, issues of search platforms’ integration of AI can fall within the regulatory scope of the MSCA.[17] Article 9 prohibits self-preferencing in search results by designated providers of search services using search engines. In the case that a designated search engine provider artificially places something produced by generative AI at the top of its query results featuring its own products or services, Article 9 can address potential problems of self-preferencing.

The JFTC will continue to monitor trends in AI and consider how to position these new technologies under the MSCA. As the technology becomes increasingly pertinent to our economy and society, the MSCA’s triennial review or updates of its guidelines and subordinate regulations may be required for further capturing the technologies if there is a lack of competition and a need for appropriate intervention.

6. Summary

As mentioned above, Japan’s current AI legal framework essentially adopts a sector-specific approach based on existing individual laws and soft law measures, representing a different path from the EU’s hard-law-based approach. The JFTC’s stance and practices are no exception to Japan’s overall approach. That said, although the primary approach will be to address issues through the traditional Antimonopoly Act framework when competition issues arise in AI markets, ex-ante competition regulations are not off the table for exceptional cases.

At present, there seems to be no clear concrete competition issues related to AI under traditional understandings of antitrust in Japan. The JFTC recognizes that in addition to the points raised in the market study’s discussion paper, there is a significant possibility that unforeseen issues may arise, however. Some of these issues may be resolved within the framework of the MSCA, while others may emerge that are difficult to address within its scope. Given the astonishing pace of development in the AI field, some problems may not be suitable for individual responses under Japan’s traditional Antimonopoly Act as well.

Following the implementation and evaluation of Japan’s AI Bill, legislators may actively consider new measures for preemptive competition regulation to complement its framework if necessary. The JFTC’s priority for now should be to stay vigilant, ensuring that innovation continues unhindered, while heeding the intentions of the basic policies that will be delivered through the AI Bill. Much like how Japan’s proposed AI legislation requires consistent monitoring of the market and active dialogue with stakeholders, the JFTC’s generative AI market study and MSCA enforcement will require robust communication – perhaps a hallmark of Japan’s new approach in tech regulation. The JFTC will strive to lead the way in this new era of regulatory balancing of safeguards and innovation so that AI can prosper.

Kyohei Yamamoto and Yasunori Tabei

Citation: Kyohei Yamamoto and Yasunori Tabei, Japanese AI Regulation and Competition Law, The Law & Technology & Economics of AI (ed. Adrian Kuenzler, Thibault Schrepel & Volker Stocker), Network Law Review, Summer 2025.

References:

• [1] AI Institutional Study Group (This website is available in Japanese only)
• [2] We refer to the meeting summaries and materials available on the AI Institutional Study Group’s website cited in footnote 2
• [3] The Interim Report (Provisional Translation) jointly published by the AI Strategic Council and the AI Institutional Study Group on February 4, 2025
• [4] Page 7 to 8 of the Interim Report
• [5] Page 9 to 12 of the Interim Report
• [6] The “General Understanding on AI and Copyright in Japan” (This document is available in Japanese only.) You can see an overview version in English here
• [7] Page 14 to 15 of the Interim Report
• [8] Page 15 of the Interim Report
• ^[9] A concept paper from the ruling Liberal Democratic Party in February 2024 and a white paper by the Cabinet Office’s AI Strategy Team in May 2024 both recommended the introduction of new regulations for large foundation models.
• [10] The bill on the Promotion of R&D and Utilization of Artificial Intelligence-related Technologies (Only Japanese version available)
• [11] Requests for Information and Comments Concerning Generative AI and Competition | Japan Fair Trade Commission
• ^[12] Policymaking and advocacy for other forms of AI are under the remit of the same division (Office of Policy Planning and Research for Digital Markets)
• ^[13] Generative AI and Competition (Discussion Paper)
• ^[14] Note that the JFTC explains how cloud computing stretches from the infrastructure layer to rollout of services; see complete report for details
• [15] Considering that the actual number of smartphone users in Japan is approximately 100 million based on thorough market research, 40 million users roughly equates to 40% of users. On the other hand, the DMA’s designation criteria is 45 million monthly active end users in the Union, and the number is equivalent to approximately 10% of the total population of the EU (448.38 million people in 2023)
• ^[16] A note for the reader that the subordinate regulations that articulate how the MSCA are to be fully enforced were released as drafts on May 15, 2025 for public comments (here)
• [17] On the other hand, if an AI tool integrates a search service, it is considered that it will not fall within the scope of the MSCA unless the AI ​​tool falls within the MSCA’s definition of a search service