Recently, AI is expected to bring significant progress to our human society. We believe that AI will enrich our lives by making our work more efficient and by increasing productivity in various industries. With AI technologies, people will be able to shift from dangerous or heavy work to more productive and creative pursuits. Moreover, a number of social problems such as labor shortage at rural farms or traffic accidents caused by human error may be solved with proactive utilization of intelligent automated technologies.
In this respect, several countries have built their own AI strategies. Germany announced its AI strategy based on the concept of industry 4.0. France also declared an AI strategy to boost AI research in France. While, in the U.S., GAFA and other private sectors are leading R&D and social implementation of AI, the White House has begun discussions on policies to take a leading role in AI. China’s AI strategy sets a goal to become the world leader of AI industries. Singapore’s AI strategy calls for a national program to gather human resources and to focus on STEM education.
On the other hand, many people have been increasing their concern about AI. For example, questions are raised such as: Will AI take our human’s job? Who will take responsibility in case of accident? Will we be controlled by AI? Will my privacy be infringed through the use of AI? When the Internet was in its earliest stages, developers did not fully think about viruses, malware, hacking, or other usage of the Internet for malicious purposes. Since AI will be incorporated into the social infrastructure, policymakers have to recognize potential risks on society which will be caused by AI.
From these viewpoints, remedies are necessary to minimize the risks as well as maximize the benefits of AI, and national AI strategies should include steps to cope with these problems as the agenda. In other words, promotion of AI technologies needs to be met with some disciplines or rules to enhance proper use of AI.
In this context, Japan has been contributing to discussions on these issues among the international society since April 2016, when Japan hosted the G7 ICT Ministers’ Meeting. Japan as the host nation proposed a draft of principles of AI R&D, over which the relevant ministers exchanged their views. As a result, the G7 countries agreed that they will continue to lead the discussions on “AI R&D Principles,” with the cooperation of international organizations such as OECD.
Within the Japanese government, MIC (Ministry of Internal Affairs and Communications) has been taking an initiative on this issue. MIC set up an expert group regarding “AI Network Society” in February 2016, and proposed a draft to be discussed at G7 as explained above.
The expert group has been continuing extensive studies, and completed “Draft AI R&D Guidelines for International Discussions” in July 2017 and “Draft AI Utilization Principles” in July 2018.
“Draft AI R&D Guidelines” is prepared as a basis for international discussions regarding matters expected to be considered in R&D activities. Taking into account that AI technologies are in the middle of development, it is not appropriate to treat the guidelines as aimed for introduction of regulations. Rather, this draft is drawn up as a proposal of guidelines that will be internationally shared as non-regulatory and non-binding soft law.
In addition, outputs or programs of AI systems might continuously change as a result of deep learning in the process of actual use. Therefore, there are other matters that AI users, not only developers, are expected to consider. Considering this aspect, the expert group further discussed on the user side on the basis of the scenario analysis (use case assessment) and ecosystem prospects formed with the progress of AI. Thus, “Draft AI Utilization Principles” compiles the basics for different types of users to be aware of to sustain the proper utilization of AI technologies.
It is expected that the deliberations on such guidelines or principles will accelerate the participation of multi-stakeholders involved in R&D and utilization of AI (such as developers, service providers, users including civil society, governments, and international organizations) at both national and international levels, in the discussions towards building “AI governance.” It also promotes the international sharing of best practices, which will help gain the trust in AI of users and the society and proper facilitation of AI.
Discussion of AI-related issues in the Japanese government is currently moving into the next phase. The Cabinet Office started a task force in September 2018 to establish Japan’s AI strategy package, and has been elaborating the details towards a human-centric AI society, incorporating united measures of all ministries and agencies. A draft of this package was disclosed in December 2018, and will be finally announced in April 2019. The five key points of this package are: 1) human resource development; 2) redesign of social systems for AI implementation; 3) data exchange platform; 4) innovation environment; and 5) social principles. The fifth is so-called ELSI (ethical, legal and social issues) and has been constructed based on “Draft AI R&D Guidelines” and “Draft AI Utilization Principles” MIC proposed.
By sharing the concept of this AI strategy package, we would like to cooperate and collaborate with international society, including the G20 summit to be held in Japan in June 2019, towards a good AI governance, that is, an appropriate balance of promotion and disciplines of AI technologies.
Finally, I would like to refer to the necessity of proceeding internationally with academic research regarding social governance or disciplines facing this digital world. These are challenges to human beings to develop into a brand-new era, and our society needs to be changed in so many ways to establish a sustainable eco-system.