Artificial Intelligence and U.S.-Japan Relations
Erik M. Jacobs (Hyogo, 2013-2016)
National security and economic issues have dominated the United States-Japan relationship in the postwar era. Science and technology (S&T) policy has received much less public attention, but it has played an important role in U.S.-Japan relations. In recent years, artificial intelligence (AI) has quickly emerged as an S&T focus for both national security and economic policy in the United States and Japan. AI, and other emerging technology issues, are likely to play a larger role in U.S.-Japan relations in the years to come.
Introduction to the History of U.S.-Japan Science and Technology Policy Relations
Science and technology policy has played a prominent role in multiple areas of U.S.-Japan relations at different points in time. Government officials, bureaucrats, and diplomats have addressed challenges regarding the scope of technological cooperation, first on national security- and space-specific applications in the 1960s, and then in crucial areas of economic competition in the 1980s, such as the semiconductor industry during Japan’s period of high economic growth. As the U.S.-Japan economic competition reached its zenith in the 1980s, U.S. President Ronald Reagan and Japanese Prime Minister Noboru Takeshita signed the U.S.-Japan Science and Technology Agreement in 1988 to identify common ground and govern S&T cooperation and collaboration between both countries. The agreement established the Joint High-Level Committee (JHLC) to address specific S&T cooperation efforts. It still guides much of the S&T cooperation between the United States and Japan, including on matters of artificial intelligence.
The JHLC mechanism has been used to address AI in various ways across American presidential administrations and Japanese prime ministers’ governments. For example, at the Fourteenth JHLC, held during the Trump administration, the JHLC split into science-specific and technology-specific tracks. Then-Deputy U.S. Chief Technology Officer Michael Kratsios led the meeting’s technology track, where advancement of AI research and development and the principles and ethics in the social implementation of AI were mentioned for the first time in the meeting readout. The Biden administration removed the two-tracked nature of the JHLC and reverted these AI-related dialogues to more S&T research-focused questions, which included U.S.-Japan cooperation on AI. Through the JHLC and other international forums, American and Japanese leaders have cooperated on various policy measures at the international level, but their internal approaches to AI have often differed.
The United States’ Approach to AI Policymaking
The U.S. government has been developing its approach to AI for several years, beginning in earnest after the Trump administration established the American AI Initiative. This initiative directed federal agencies to report all of their known nonsensitive and nonclassified uses of AI to support government efforts to track AI use cases, including research investment, federal computing, technical standards, the AI workforce, and engaging with allies on artificial intelligence. These processes were eventually codified with the passage of the National AI Initiative Act of 2020. President Trump also signed his executive order on Promoting the Use of Trustworthy AI in the Federal Government in December 2020, which first defined principles for the use of AI in government, established implementing principles, and directed agencies to catalog AI use cases, among other things.
In October 2023, the Biden administration announced the executive order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, which outside observers have said will be the cornerstone for the Biden administration’s regulatory approach to AI policy. This executive order has created eight different principles and priorities to guide the Biden administration’s approach to AI. It aims to govern what the administration identifies as AI safety, innovation, AI development, the use of AI to advance the administration’s view of equity and civil rights, interaction with AI, privacy and civil liberty protections, risk management, and AI’s societal impacts. Of note, the application of these priorities is not limited to generative AI and is expected to address wider implications of AI policy across the government.
At the agency level, the executive order directs the National Institute of Standards and Technology (NIST) to develop guidelines on developing and deploying “safe, secure, and trustworthy AI systems” and standards for internal security tests to identify AI vulnerabilities with developers. The order directs the Department of Energy to develop evaluation tools for its use of AI in applications related to chemical, biological, nuclear, and cybersecurity risks, as well as AI systems’ threats to critical infrastructure. The executive order also directs the Department of Commerce to establish requirements for companies seeking to pursue AI dual use models within ninety days, directs the Office of Management and Budget to establish a council on the use of AI in the federal government, and directs the Department of State to identify ways to streamline processing times of visa petitions for foreign nationals seeking to work on AI technologies in the United States. Finally, the executive order directs the Secretaries of Commerce and State to establish a plan for global engagement with international partners on AI terminology, best practices, trustworthiness, and risk management. Such engagement could present an opportunity for short-term cooperation between the United States and Japan on AI-related issues.
On the legislative front, Congressional Democrats have focused on similar issues by introducing legislation such as the Algorithmic Accountability Act of 2023 that would expand the regulatory powers of the Federal Trade Commission, but it is unlikely that sweeping AI legislation will be enacted with split control of the 118th Congress.
Japan’s Approach to AI Policymaking
For several years, Japanese policymakers have also sought to address AI with an emphasis on societal impact while also taking measures to prevent overcompensation for AI’s perceived risks. These efforts gained momentum in 2019 when the Japanese government released its Social Principles of Human-Centric AI, which lays out the framework for how the Japanese government would approach AI-related policy issues. When addressing emerging policy issues on trade and in the digital space, Japan has traditionally favored engagement in line with multilateral organizations rather than a unilateral approach, and the same can be said for its approach to AI. Its Social Principles for Human-Centric AI are broadly in line with the Organization for Economic Cooperation and Development (OECD)’s AI Principles, which were first released in May 2019.
Importantly, Japan’s approach to AI development is to ensure that AI technologies contribute to the realization of its so-called “Society 5.0” objectives, which were first proposed in Japan’s Fifth Science and Technology Basic Plan. Through its AI development strategy, Japan aims to create an “AI-ready society” that is shaped by three basic philosophies: dignity, diversity and inclusion, and sustainability. The government has also laid out several social principles that should be implemented across Japanese society to create this AI-ready society: the human-centric principle; the principle of education/literacy; the principle of privacy protection; the principle of ensuring security; the principle of fair competition; the principle of fairness, accountability, and transparency; and the principle of innovation.
At the time of publication, Japan does not have an overarching regulatory structure or legislation that governs the entire government’s use of AI despite support for the idea from Prime Minister Fumio Kishida. Ministries have begun forming policy guidelines and targeted legislation has also addressed specific AI use cases. For example, the Digital Platform Transparency Act addresses certain areas of online commerce and the Financial Instruments and Exchange Act addresses online financial activities while the Japan Fair Trade Commission uses its mandate to analyze potential algorithmic risks. In January 2022, the Ministry of Economy, Trade and Industry (METI) released governance guidelines for the implementation of Social Principles of Human-Centric AI. In its eighty-two-page report on the topic, METI presented examples of how to use Japan’s AI governance guidelines in specific use cases, with an emphasis on collaboration with stakeholders and outside groups.
Prime Minister Kishida’s government is expected to release legally non-binding AI guidelines for businesses in late 2023. These guidelines would separate businesses into several different categories: developers, businesses that train AI, systems developers that incorporate AI, service providers using AI, and users. They are also likely to call on AI platform developers to share information about the purpose of their algorithms. Under these proposed guidelines, companies using AI will also have to take measures to reduce overreliance on AI, not use biased data, and maintain records of the AI technology. These possible regulations come amid reports that the Kishida Cabinet will request 168 billion yen ($1.11 billion) for AI-related spending in the fiscal year 2024 budget.
The Japanese government has also attempted to position itself as a world leader in AI regulation during its year-long term holding the Group of 7 (G7) Presidency in 2023. The most notable development through this process is the G7 Hiroshima AI Process, which aims to address the opportunities and potential of AI systems, foundation models, and generative AI. Japan’s vision for the G7 Hiroshima AI Process is based on the multi-stakeholder consultation process and cooperation with the OECD and the Global Partnership for Artificial Intelligence. In October 2023, G7 leaders agreed on a voluntary code of conduct at the international level, advancing a key AI priority for the Kishida government as it seeks to influence AI policy throughout multilateral institutions.
While the United States and Japan share similar policy goals, both countries have adopted slightly different approaches to the implementation of regulatory frameworks. Japan has taken a lighter-touch approach towards regulation to avoid excessive restrictions on the burgeoning AI industry while the Biden administration has pursued a broader regulatory approach. The Biden administration has faced criticism about its recent executive order for potentially stifling innovation across the AI sector from industry groups for this approach, and former U.S. Chief Technology Officer Michael Kratsios has warned about advancing a regulatory regime without clear understandings of potential AI use cases in agencies or assessments of where AI will potentially impact a new regulatory regime. Despite these differing domestic approaches to AI, American and Japanese policymakers have adapted several different workstreams within government to drive bilateral cooperation, primarily at the ministerial level.
State of U.S.-Japan AI Cooperation
At this stage in U.S.-Japan cooperation on civil AI issues, METI and the U.S. Department of Commerce lead much of the work through various workstreams. For example, the Japan-U.S. Economic Policy Consultative Committee (Economic 2+2) includes cooperation on AI to strengthen economic resilience and promote and protect critical and emerging technologies for both countries. As recently as November 2023, the Biden administration has said that it supports the G7 Hiroshima AI Process as well as two additional processes: (a) the Hiroshima Process International Guiding Principles for Organizations Developing Advanced AI Systems and (b) the Hiroshima Process International Code of Conduct for Organizations Developing Advanced AI Systems. An official statement from the Economic 2+2 also outlined other cooperation on AI guidelines between Japan’s Information-Technology Promotion Agency and the NIST AI Risk Management and the development of semiconductors that are necessary for the development of generative AI. Other bilateral cooperation exists between the U.S. Department of Commerce and METI through the Japan-U.S. Commercial and Industrial Partnership (JUCIP), but at a smaller scale than the Economic 2+2 dialogue. There has also been cooperation between the U.S. Department of Defense and Japan’s Ministry of Defense on the responsible military use of AI.
Outside of the G7 and OECD, the United States and Japan have cooperated on AI-related topics through technical standards groups within the structure of the Quadrilateral Strategic Security Dialogue (Quad), which includes the United States, Japan, Australia, and India. There has also been limited cooperation with other like minded countries, such as South Korea. Additional cooperation efforts within larger international organizations, such as the Asia-Pacific Economic Cooperation (APEC), may be challenging and also may not be advisable, as evidenced by the most recent Leaders’ Summit failing to address AI-related cooperation.
Conclusion: The Future of U.S.-Japan AI Cooperation
While the United States and Japan are pursuing slightly different domestic approaches to AI regulation and policymaking, current governments in both countries are taking similar approaches to engagement on AI-related policy issues at the international level. Opportunities for enhanced government-to-government cooperation may be limited to high-level JHLC coordination and specific ministerial-level efforts between the U.S. Department of Commerce and METI, NIST and Japan’s Information-Technology Promotion Agency, and the Pentagon and Ministry of Defense as both countries figure out their approaches to the burgeoning AI area. Given the key role private industry plays in emerging technology, deeper private sector coordination will also be essential for the future of U.S.-Japan AI cooperation. Such coordination is especially true when it comes to the development of next-generation semiconductors to enhance both economic security and national security as Washington and Tokyo seek to build secure supply chains that do not rely on Chinese components. The Biden administration should pursue a lighter approach to regulation of private industry while it figures out how AI will impact specific industries, as the risks of stifling innovation with both executive order and poor policy are high. The current governments in Washington and Tokyo are likely to maintain current AI-related bilateral efforts while also seeking cooperation through international organizations and multilateral frameworks such as the G7 process while global AI regulatory regimes are in their nascent stages. Regardless of short-term policy actions, AI is set to play a major role in certain parts of commercial diplomacy and national security relations within the U.S.-Japan alliance for years to come.
About the Author
Erik M. Jacobs taught English as an Assistant Language Teacher on the Japan Exchange Teaching (JET) Program from 2013 to 2016 in the coastal city of Kobe-shi, Hyogo Prefecture. After returning from Japan, Erik received an MA in Asian Studies from Georgetown University and has been involved in the U.S.-Japan relationship in various government roles. He served as the project manager for the ministerial level Fourteenth Joint High-Level Committee Meeting on Science and Technology Cooperation Between the United States and Japan at the White House in 2018 and worked on two Vice Presidential visits to Japan in 2018. Erik continues to write on various components of U.S.-Japan relations with a focus on science and technology cooperation. Currently, he is a policy manager for U.S. and Asia at Access Partnership and is a nonresident scholar with the Sasakawa Peace Foundation USA. Erik holds a BA in political science from Temple University with a concentration in Japanese and was a study abroad student at Temple University, Japan Campus in Tokyo during the March 11, 2011, Great Tohoku Earthquake.
JETs on Japan is a partnership between USJETAA and Sasakawa Peace Foundation USA (Sasakawa USA) that features selected articles of JET alumni perspectives on US-Japan relations. The series aims to elevate the awareness and visibility of JET alumni working across diverse sectors and provides a platform for JET alumni to contribute to a deeper understanding of US-Japan relations from their fields. The articles will be posted on USJETAA’s website to serve as a resource to the wider JET alumni and US-Japan communities on how alumni of this exchange program are continuing to serve as informal ambassadors in US-Japan relations.
Submissions are encouraged from mid-to-senior level professionals who are established in the current fields OR current/recent graduate degree students in both master’s and doctoral programs.
