人工智能如何改变亚洲的保险业

Table of ContentsPreface................................................................................................................................... 2Definition of AI and Scope...................................................................................................... 61.Purpose........................................................................................................................... 72.Basic Philosophies.......................................................................................................... 93.Classification of Related Entities................................................................................... 104.AI Utilization Principles...................................................................................................115.General Flow of AI Utilization........................................................................................ 136.Commentary on the AI Utilization Principles................................................................. 157.Timing to Consider the AI Utilization Principles ............................................................ 28(Attachment)Detailed Explanation on Key Points Concerning AI Utilization Principles 1 PrefaceResearch and development (R&D) and utilization of artificial intelligence (AI)are expected to progress dramatically in the years to come, and discussions areongoing internationally to foster trust in AI. In Japan, from the viewpoint ofpromoting the transformation to an “AI-ready society,” the government startedstudying common AI social principles in May 2018 and formulated the “SocialPrinciples of Human-Centric AI”1in March 2019.At the G7 ICT Ministers’ Meeting in April 2016, Japan as the host nationintroducedthe principles of AI development,which triggered a study onAI-related principles. At the G7 ICT Ministers’ Meeting, discussions were held byrelevant ministers, and the G7 countries agreed to continue having a central roleand to discuss formulating the “AI R&D Guidelines”, which consist of “AI R&Dprinciples” and explanations for them, with the cooperation of internationalorganizations,such as the Organization for Economic Co-operation andDevelopment (OECD). For the promotion of the benefits of AI with risk mitigation,theConference toward AI Network Society (hereinafter referred to as the“Conference”) studied items that should be taken into consideration at the timeofresearch and development.In July 2017,the Conference released adocument entitled “AI R&D Guidelines for International Discussions”(hereinafterreferredto as the“Draft AI R&D Guidelines”)to serve as the basis ofinternational discussions in the G7 and the OECD.On the other hand, AI may change its implementation and output continuouslyby learning from data in the process of its utilization. Therefore, it is assumedthat there are not only matters that developers are expected to take intoconsideration,but also matters that users 2 are expected to take intoconsideration when using AI. Furthermore, it is considered important to itemizematters to be taken into consideration in the utilization of AI from the viewpoint ofstudying roles expected from various stakeholders, including developers, users,and data providers. For this reason, the Conference released the “Draft AIUtilization Principles” in July 2018 as rules to which AI service providers, end1“Human-centered AI Social Principles” (decision by the Council for Integrated InnovationStrategy In Japan in March 2019)2Refer chapter 3 regarding classification of related entities such as “developers” and“users”, etc. 2 3users, and data providers are expected to pay attention.Japan has been actively disseminating the study results to the internationalcommunity. In particular, at the meeting of the AI expert group at the OECD(AIGO), which was established by the OECD in September 2018 to draft therecommendations of the OECD Board, Japan made a significant contributionthat included the introduction of Japan’s “Social Principles of Human-CentricAI”, ”Draft AI R&D Guidelines”, and “Draft AI Utilization Principles” . Japaneseexperts introduced not only the contents of these principles and guidelines, butalso the background of the study and situation of discussion that took place onthem. For this reason, the Recommendation on Artificial Intelligence of theOECD Board, published in May 2019, is consistent with the above principles andguidelines that have been discussed in Japan.Furthermore, the “G20 AI Principles”, which adopted the recommendations oftheOECD Board,were compiled at the G20 Trade and Digital EconomyMinisters Meeting (June 2019), at which Japan served as the host country.As described above, Japan has led international discussions on the principlesof AI. As a result, international consensus has been built on the concept of theprinciples. Meanwhile, progress continues on studying specific measures thatare expected to be needed in the future, such as measures to realize theprincip