Japan: Bill on promoting research, development and utilization of artificial intelligence-related technologies
- Ksenia Laputko
- 1 day ago
- 4 min read
Remember when I said we should expect an explosion of AI legal frameworks in the coming years? Well, here’s another to add to the growing collection: Japan has just introduced the “Bill on Promoting Research, Development and Utilization of Artificial Intelligence-Related Technologies.”
Let’s unpack it briefly.
Structure
The bill is structured as follows:
Chapter I: General Provisions (Articles 1–10)
Chapter II: Basic Measures (Articles 11–17)
Chapter III: Basic Plan for AI (Article 18)
Chapter IV: AI Strategy Headquarters (Articles 19–28)
It supplements Japan’s Basic Act on Science and Technology and Innovation (1995) and the Basic Act on the Formation of a Digital Society (2021).
Definition of AI. Functional, but lacks precision
The bill defines “AI-related technology” as technologies that simulate human cognitive functions (like reasoning and judgment), as well as systems that process and output data using such technologies. It’s a broad and functional definition, more aligned with general engineering perspectives than with legal clarity.
Compare this to the EU AI Act, which defines AI based on core characteristics like autonomy and adaptability—a more structured approach that aligns well with legal risk frameworks.
Strategic focus or lack of farsight?
The bill emphasizes national competitiveness, economic development, and administrative efficiency. Its tone and priorities align closely with the U.S. policy direction under the Biden administration—encouraging innovation first, and regulating later.
A central provision states:
“The State shall promote the active use of artificial intelligence-related technologies in national administrative agencies in order to improve the efficiency and sophistication of administrative affairs.”
This provision raises serious red flags. Why? Because Japan’s Act on the Protection of Personal Information (APPI) does not include explicit provisions on automated decision-making with legal or similarly significant effects—unlike the GDPR or the EU AI Act.
Without those safeguards, the use of AI in public administration could lead to unaccountable, opaque decision-making, undermining human rights and procedural fairness.
What’s missing? Risk, accountability, and developer duties
One of the most striking gaps is the bill’s complete absence of obligations for AI developers, deployers, or service providers. Most of the measures are addressed to the state—such as promoting research, facilitating commercialization, or sharing R&D results.
"Article 7 Any person who intends to develop or provide products or services that utilize artificial intelligence-related technology, or any other person who intends to utilize artificial intelligence-related technology in their business activities (hereinafter referred to as "Utilizing Businesses") must endeavor to improve the efficiency and sophistication of their business activities and create new industries through the active utilization of artificial intelligence-related technology in accordance with the Basic Principles, and must cooperate with the measures implemented by the national government pursuant to the provisions of Article 4 and the measures implemented by local governments pursuant to the provisions of Article 5.
And thats literally it.
Ah, wait, here is "AI literacy":
"Article 8 The public shall, in accordance with the Basic Principles, endeavor to deepen their understanding of and interest in artificial intelligence-related technologies and to cooperate with the policies implemented by the national government pursuant to the provisions of Article 4 and the policies implemented by local governments pursuant to the provisions of Article 5."
AI strategy headquarters
There’s also an ambitious plan to set up an AI Strategy Headquarters, but oddly enough, the bill goes into great detail about its composition, while saying very little about its actual powers, functions, or oversight mechanisms.
Again here is the provison:
"In order to comprehensively and systematically advance policies concerning the promotion of research, development, and utilization of artificial intelligence-related technology, an Artificial Intelligence Strategy Headquarters (hereinafter referred to as the "Headquarters") shall be established within the Cabinet."
Japan’s bill is strategic and forward-looking—but decidedly one-sided. It reflects a national ambition to lead in AI innovation, but neglects to address the very real risks that come with widespread deployment, particularly in public administration.
In contrast to the EU AI Act, which embraces a risk-based model rooted in fundamental rights, Japan’s legislation is closer to industrial policy than rights-based governance. And while innovation matters, without robust safeguards, it risks eroding public trust—something no AI strategy can afford.
Want to know more about AI or privacy? Preparing for AIGP of CIPP exams? Don't miss my courses and books:
Buy my author approach prep books on Amazon:
1. CANADA DATA PROTECTION LAW (fit for CIPP/C Study Guide) https://a.co/d/drVur9r
2. European Data Protection Law (fit for CIPP/e prep) https://a.co/d/2DxUL2J
3.CIPP/A & CIPP/CN Study Guide 2024: Hong Kong, Singapore, India, China data privacy frameworks https://a.co/d/dUvKinY
4. AIGP Study guide: AI Governance from scratch https://a.co/d/dKbgT4j
Courses with May %
1) #cippc Canada data privacy law author prep https://lnkd.in/e76qUb7Y
2) #cippe European data protection law:author prep https://lnkd.in/eVuGDw_r
3)#cippa Asia (India, Singapore, Hong Kong) Privacy :author prep https://lnkd.in/ehcX4Dp2
4)#cippus USA privacy https://lnkd.in/ekTSGsyy
Disclaimer: This unofficial study guides represent my analysis and approach. This books have not been developed in conjunction with the IAPP, nor is it endorsed, sponsored, funded, or supported by the IAPP. It does not represent the official IAPP position or study materials.