Study Group on Issues and Policies in the Use of AI

In recent years, AI technology has made groundbreaking progress and is expected to bring opportunities for significant social change, but there are also many issues to be addressed, including ethics, copyright, false information, and security. In order to discuss a wide range of policies and issues that need to be addressed to promote new innovation and efficiency through the use of AI in order to realize a better digital society, and to disseminate information and make proposals, the "AI Study Group" was newly The group has been launched.

With the use of AI, discussions on regulations for AI have been in full swing in many parts of the world. In Japan, too, there is no dispute that some legal discipline should apply to the use of AI at least in existing regulatory fields (e.g., medical, financial, transportation, infrastructure, consumer protection, etc.), and thus, laws are being actively revised to enable AI use in regulatory fields (e.g., Pharmaceutical Affairs Law, Financial Instruments and Exchange Law, Road Traffic Law, Digital Procedures Law, High Pressure Gas Safety Law, Installment Sales Law, etc.). However, discussions on what specific conditions must be met for AI to meet the statutory standards are still in the middle of the road. Regulations are also being discussed in various countries, including Japan, regarding generative AI, which has a large social impact but for which the full extent of risks has not yet been clarified.

One theme of the study group is for each of the following areas: healthcare, finance, transportation, infrastructure, consumer protection, and generative AI,
(1) What AI systems will require certification or licensing?
(2) What should certification and licenses be granted, to whom, in what process, and for what purpose?
Based on the above research and analysis, we are considering the following issues (Phase 1 of the research): 1) analyzing the systems and trends of discussions in each country (Phase 1 of the research), 2) drawing up an overall picture of how Japan should utilize AI (what can be applied across fields and to what extent, and what needs to be specialized in each field) and whether cooperation among different countries is feasible (Phase 2 of the research). We are considering drawing up an overall picture (Phase 2 of the research) of how Japan should apply AI (to what extent it can be applied across fields, and what needs to be specialized in each field) and the possibility of cooperation among different nations.

Lead and participating committee members　(Titles omitted. Positions are as of April 2024, when the Study Group was established)

Chair:
Hiroki Habuka
(Research Professor, Graduate School of Law, Kyoto University / Attorney)
Committee Member:
Tatsuhiko Inatani
(Professor, Graduate School of Law, Kyoto University)
Takafumi Ochiai
(Attorney at Law (Senior Partner), Atsumi & Sakai / Co-Founder and Representative Director, Smart Governance Inc.)
Ryoichi Sugimura
(Chief Collaboration Officer / National Institute of Advanced Industrial Science and Technology (AIST), Information and Ergonomics Area)
Kumiko Takahashi
(Consultant, Risk Management Group, Social Infrastructure Business Division, Mitsubishi Research Institute, Inc.)
Keisuke Tomiyasu
(Deputy Director, Behavior Informatics Laboratory, FRONTEO Inc.)
Keita Nishiyama
(Visiting Professor, Institute for Future Initiatives of the University of Tokyo)

[Related information]
　How to design a certification system for AI? [Japanese ]
　Hiroki Habuka, Research Professor, Graduate School of Law, Kyoto University

[Commissioned report of the study group]
　Comparative Legal Study on AI System Certification [English ][Japanese ]

1st meeting
April 1, 2024 13:00-14:30
Mr. Mikiyuki Kato, Director of the CFIEC Institute for Digital Society, gave a presentation on the "Inauguration of the CFIEC Institute for Digital Society as of April 1," followed by a question-and-answer session on a proposal by Mr. Hiroki Habuka to conduct a survey on trends in AI certification and licensing in various foreign countries. The participating committee members expressed their expectations for the implementation of the survey and also commented on the direction of the selection of countries and fields to be surveyed, as well as on the need to delve deeper into the background and hidden strategies based on the results of the analysis of the survey results.

2nd meeting
June 12, 2024 15:30-17:00
The secretariat explained about the general trends in global AI legislation and systems that occurred after the first meeting (April 1), and also explained about the "Overview of Recent Legal Amendments to Promote AI in Japan" based on the commissioned survey on trends in AI certification and licensing in various countries and industrial sectors, which was endorsed by the participating members at the first meeting. The secretariat also gave an overview of recent legal amendments to promote AI in Japan, based on the survey on trends in AI certification and licensing in foreign countries and industry sectors, which was endorsed by the participating members of the first meeting. Then, Mr. Ryoichi Sugimura, Chief Collaboration Officer, National Institute of Advanced Industrial Science and Technology (AIST), made a presentation on "AI International Standardization and Major Movements in the World". The participating committee members expressed their expectations for the continuation of the trend survey as well as their opinions on the promotion of international standardization strategies for AI.

3rd meeting
August 6, 2024 13:00-15:00
Mr. Habuka, who was in charge of the survey, Mr. David Socol de la Osa David Uriel, Associate Professor at Hitotsubashi University, and Mr. Kei Takahashi, Attorney at Law, made an interim presentation on the "Survey on AI Certification and Licensing Trends in Other Countries". The objectives of AI laws in the EU and China and the differences in their registration systems were discussed. It was also suggested that the so-called joint certification discussed in ISO, i.e., a combination of organizational/process certification and product certification, might be helpful when considering the introduction of an AI licensing system in Japan. In the next meeting, a deeper investigation of trends based on the points raised at the 3rd meeting will be conducted.

4th meeting
October 2, 2024 12:30-14:30
Continuing from the previous meeting, Dr. David Socol de la Osa David Uriel, Associate Professor at Hitotsubashi University, and Mr. Kei Takahashi, Attorney at Law, presented an interim report on the licensing and certification systems for automated driving vehicles and medical devices in Japan, the UK, and the US, which are relatively advanced in terms of AI systems and AI models. Based on this report, the following issues were discussed: (1) the concept of allowing low-level traffic violations by automated vehicles and the concept of corporate penalties under the UK Automated Driving Act, (2) the ideal form of organizational certification when AI products and organizations that develop and manufacture such products are certified together, (3) standards to be followed by high-risk AI in the EU AI Act and the relationship with ISO. (3) the relationship between the standards to be followed by high-risk AI under the EU AI Law and ISO.

5th Meeting
Thursday, December 5, 2024, 10:30-12:30
A report was presented on the almost final report of the comparative analysis of the certification systems for AI systems in each country, covering automated driving vehicles and medical devices. The main findings of the report were: (1) the system design is still under development in each country; and (2) there is still room for further study on the combination of products, management systems, and operators, which are the components of joint certification. The following issues were also discussed: 1) institutional framework for certification of AIs, 2) methods of certification, 3) the nature of certification bodies, 4) the relationship between certification and liability systems, and 5) how to identify risks as a precondition for the design of certification systems.

6th Meeting
Friday, March 21, 2025, 9:30-11:30
An expert in charge of certification at a U.S. company explained the challenges of conformity assessment for AI systems and joint certification as a solution to these challenges. The workshop agreed that, based on the results of the FY2024 study group, the ideal form of joint certification should continue to be considered in the new year.