• 2023.10.31 ｜

  Policy Recommendation

  ｜

  No.25

  Technology Governance Policy Research Unit

International Collaboration in AI Governance :
Key Considerations of the Council of Europe’s AI Convention and Japan’s Response

Currently, artificial intelligence (AI) technologies are widely deployed with high expectations of transformative impacts on daily life and work. Consequently, owing to its significant impact it may have on society, AI governance has become a prominent subject of discussion in international organisations, and networks, including the G7 Hiroshima Summit. The Hiroshima Summit mandated the creation of the G7 Hiroshima AI Process to harmonise various regulatory frameworks and promote AI governance in a coordinated manner, particularly with regard to generative AI. In addition, the Ministerial Declaration of the G7 Digital and Tech Ministers’ Meeting acknowledges the need for agile and decentralized governance with multi-stakeholder participation, and legal frameworks, designed to operationalise the principles of the rule of law, due process, democracy, and respect for human rights.

The Council of Europe (CoE) is engaging in discussions concerning the Convention on AI, which considers values such as the rule of law, human rights, and democracy in the life cycle of AI, from design to decommission. Negotiations are currently underway to draft the world’s first convention on AI, which includes non-member states participating as observer states. A framework convention, as such, while possessing legally binding authority, grants contracting parties autonomy to determine the means to achieve convention objectives. Key provisions include taking measures to ensure the availability of effective remedies for violations of human rights and fundamental freedoms, record-keeping for such incidents, and the implementation of risk and impact assessments based on a risk-based approach. The Convention also stipulates the establishment of oversight mechanisms responsible for monitoring compliance.

All G7 countries participated in the drafting and negotiation of the AI Convention, and Japan, as an observer state, is in a position to influence the formation of AI-related rules. Considering this, Japan, as the chair of the G7 Hiroshima AI Process, should position the Convention as a tool for expanding AI-related rules worldwide based on the shared values of the G7 countries and be actively involved in negotiations. Given this context, this policy recommendation identifies the points that Japan should clarify its positioning in view of the negotiation process toward the conclusion of this Convention and proposes the following three points.

Many provisions make it difficult to determine whether new legislation is required to align with this Convention or whether existing laws, regulations, and guidelines can be addressed without new legislation. In Japan, in particular, horizontal regulations on the development and use of AI systems have been addressed by guidelines, rather than legislation, and existing domestic laws pertaining to AI are limited. In addition, in Japan, it is necessary to examine how regulation should be applied to AI use not only in the private sector but also in the public sector, including the government. Furthermore, as the discussion on use of AI in defence and national security advances globally, discussions should be organised domestically, while taking account of the economic security policy perspectives. As an observer state, Japan should promptly confirm the existence of relevant domestic laws and guidelines, clarify desirable directions, and actively participate in drafting and negotiating the Convention.

The present draft content of the Convention stipulates that measures should be taken for risk impact assessments to be conducted with respect to AI systems within the scope of the Convention. Therefore, a risk and impact assessment methodology and system that are feasible for relevant organisations should be developed in Japan as soon as possible, while considering trends in international standard risk and impact assessment frameworks. Currently, the Japanese government is in the process of formulating “Guidelines for AI Businesses”. While the contents of the Guidelines and this Convention need not be identical, the aim should be to ensure that there are no contradictions and that information on the Japanese Guidelines should be advocated.

The Convention is envisioned to mandate the establishment of an independent and impartial mechanism with the necessary authority and resources to monitor and oversee compliance with obligations. Although designating an existing institution as a structure or to using multiple institutions as an oversight mechanism is allowed under the Convention, concrete discussions should be initiated as soon as possible on how to create an efficient structure within the government, with an eye to the future, while referring to the responses of other countries.

The oversight mechanism will protect personal rights and play a role in asserting the legitimacy of Japan’s efforts in other countries. Therefore, it is crucial to promptly consider and designate existing administrative organizations involved in the oversight mechanism and the number of personnel that are experts in AI should be increased and educated. In addition, experts should be promptly appointed to be involved in the negotiation stage of this Convention, and to hasten the establishment of a system that will receive support from the industry, academia, public, and private sectors. The main members of the oversight mechanism should be individuals who are capable of performing their duties over a long period of time with continuity, so that they can be recognised domestically and internationally as negotiating parties or contact points.