AI Guidelines for Business Ver 1.0 (PDF)

AI Business Guidelines Version 1.0 April 19, 2024 Ministry of Internal Affairs and Communications Ministry of Economy, Trade and Industry Table of Contents Introduction Part 1. What AI is Part 2. The society that AI should help realize and items each actor should address A. Basic philosophy B. Principles C. Common guidance D. Common guidance for businesses related to advanced AI systems E. Establishing AI governance Part 3. Matters concerning AI developers Part 4. Matters concerning AI providers Part 5. Matters concerning AI users Introduction AI-related technologies are developing every day, and opportunities to use AI and its various possibilities continue to expand. AI is being used to create innovation in industry and to solve social issues. With the recent rise of conversational generative AI, democratization of AI has occurred, allowing many people to easily use AI for various purposes through dialogue. As a result, companies are not only incorporating AI into business processes but are also working to reconstruct business models themselves in light of the value created by AI. Individuals are also accelerating efforts to reflect their own knowledge into AI and expand their productivity. Japan has long advocated a human-centered society that balances economic development and the solution of social issues through systems that highly integrate cyberspace and physical space as Society 5.0. To realize this concept, the Social Principles of Human-Centric AI were established in March 2019 so that AI will be accepted by society and used appropriately. As the scope of AI technology use and the number of users expand, risks are also increasing. In particular, with generative AI there are new social risks that previous AI did not pose, such as infringement of intellectual property rights and the generation and dissemination of misinformation and disinformation. The social risks brought about by AI are becoming more diverse and greater. Against this background, these Guidelines present a unified direction for AI governance in Japan so that safe and secure use of AI will be promoted. They aim to help people who use AI in various business activities correctly recognize AI risks in light of international trends and stakeholder concerns, and to voluntarily implement necessary measures across the entire AI lifecycle. Through collaboration among related parties and through practicing common guidance and the matters important to each actor together with AI governance, the Guidelines aim to proactively co-create a framework that both promotes innovation and mitigates risks across the lifecycle. Japan has led international discussions in forums such as the G7, G20, and OECD. At the same time, several points have been noted when advancing concrete implementation of AI principles: - AI use is expected as a means to solve social issues such as labor-force decline associated with a shrinking and aging population. - There can be a time lag between legal development and enforcement and the speed and complexity of AI technological development and social implementation. - Rule-based regulation that prescribes detailed conduct obligations can inhibit innovation. Based on these considerations, the Guidelines were created using a goal-based approach that seeks to achieve policy objectives through non-binding soft law while encouraging voluntary efforts by stakeholders, reducing social risks from AI, and promoting AI innovation and use. Under this understanding, the ministries had previously published the draft AI R&D Guidelines for international discussion, the AI Utilization Guidelines, and the Governance Guidelines for Implementation of AI Principles Ver. 1.1. This time, those three guidelines were integrated and reviewed to reflect the characteristics of AI technology that have further developed in recent years and discussions on AI social implementation in Japan and abroad. The result is a new set of Guidelines for business operators to implement AI social deployment and governance together as non-binding soft law. In place of the previous guidelines, these Guidelines enable businesses that utilize AI, including public institutions such as government and local governments, to confirm guiding principles for desirable action for the safe and secure use of AI. The Guidelines were formulated with emphasis on effectiveness and legitimacy through repeated discussion by a multi-stakeholder group including educational and research institutions, civil society including general consumers, and private enterprises, rather than by the government alone. Use of AI can produce large risks to society depending on the field and mode of use, and social friction caused by those risks can itself inhibit the use of AI. At the same time, excessive measures can also inhibit AI use and the benefits obtained through it. Therefore, it is important to understand in advance the magnitude of risks that can arise from the mode of use in a given field and to align the degree of countermeasures with the magnitude of risk under a risk-based approach. The Guidelines describe the direction of measures in companies based on this risk-based approach. This way of thinking is widely shared among advanced AI countries. As AI trends change rapidly, the Guidelines are expected to be updated as a Living Document under multi-stakeholder participation, taking into account international discussion and the concept of agile governance. This manual import source is derived from official METI PDF text exposed through the official web viewer for: https://www.meti.go.jp/press/2024/04/20240419004/20240419004-1.pdf