Launch of the “Reporting Framework” for the International Code of Conduct (“Hiroshima AI Process”)

The announcement says a Reporting Framework linked to the Hiroshima AI Process has been launched on the OECD website. The framework is a monitoring mechanism that lets organizations developing advanced AI systems share information on how they apply the International Code of Conduct, which they adopt voluntarily.

The announcement explains the background. In 2023, during Japan's G7 presidency, the G7 started the Hiroshima AI Process, a discussion framework on international governance of advanced AI, and approved both the International Guiding Principles and the International Code of Conduct. During Italy's G7 presidency in 2024, members worked on tools for monitoring application of the Code in order to encourage transparency and accountability in advanced AI development. With the OECD's cooperation, the G7 settled on a questionnaire and an operating procedure, and the framework went live on the basis of those two documents.

The questionnaire for AI developers, drawn up in line with the Code, is available at the OECD transparency website. As a rule, organizations that answer it will be listed on the OECD site and their responses made public. It has seven topics: risk identification and evaluation; risk management and information security; transparency reporting on advanced AI systems; organizational governance, incident management and transparency; content authentication and provenance mechanisms; research and investment to advance AI safety and reduce social risks; and advancing human and global interests.

Several Japanese AI development organizations have said they intend to take part, and a list of them accompanies the announcement. Other organizations may also join by completing the questionnaire and allowing their answers to be published on the OECD website.

In 2023, under Japan’s G7 Presidency, G7 launched the "Hiroshima AI Process," a framework for discussion regarding international governance framework for advanced AI systems, and agreed on the "International Guiding Principles" and "International Code of Conduct". Under the G7 Italian Presidency in 2024, G7 was discussing the development of the monitoring mechanism (the “Reporting Framework”) for sharing the information on the status of application to the International Code of Conduct by organizations developing advanced AI systems who voluntarily commit to it. Today, the “Reporting Framework” was launched on the OECD website.

1. Background

In 2023, under Japan’s G7 Presidency, G7 launched the "Hiroshima AI Process," a framework for discussion regarding international governance framework for advanced AI systems, and agreed on the "International Guiding Principles" and "International Code of Conduct". Under the G7 Italian Presidency in 2024, G7 discussed how G7 develop and introduce an appropriate tools and mechanisms (the “Reporting Framework”) for monitoring the application of the “International Code of Conduct”, in order to promote transparency and accountability in the development of advanced AI systems. As a result, with the cooperation of the OECD, the G7 agreed on the questionnaire and the operational procedure of the "Reporting Framework". Today, based on the approved operational procedure and the questionnaire, the “Reporting Framework” was launched on the OECD website.

2. Overview

The questionnaire for AI development organizations, elaborated in line with the “International Code of Conduct” is now made available on the OECD website (https://transparency.oecd.ai/). In principle, AI development organizations that have responded to the questionnaire will be listed on the OECD website, and their responses will also be made public. The questionnaire includes the following items: 1.Risk identification and evaluation 2.Risk management and information security 3.Transparency reporting on advanced AI systems 4.Organizational governance, incident management and transparency 5.Content authentication & provenance mechanisms 6.Research & investment to advance AI safety & mitigate social risks 7.Advancing human and global interests

3. Participating companies from Japan

The following AI development organizations from Japan have expressed their intention to participate in the “Reporting Framework”. Other organizations can also participate by completing the questionnaire and making their responses available on the OECD website.

[List of organizations from Japan that have expressed their participation] *In alphabetical order

(Reference) About the Hiroshima AI Process Please visit the website containing information about the Hiroshima AI Process: (https://www.soumu.go.jp/hiroshimaaiprocess/)