Since the end of World War II, Japan has not had a full-fledged military, but only Self-Defense Forces (SDF) for the sole purpose of defensive defense. Hesitancy exists in society, especially in academia, to research and develop technologies that could be diverted to military use. This has created a long shadow of the past in which public opinion, strategic culture, and the academic-industrial ecosystem mutually reinforce each other not to directly address defense technologies. However, with the recent deterioration of the national security environment around Japan, such hesitance is weakening, and research and development of technologies that can be applied to defense purposes is now being conducted, also in cooperation with partners like the United States, the United Kingdom, and others.

Discussions about artificial intelligence for defense purposes began in earnest around 2022, with descriptions found in government and defense ministry documents, and budget appropriations beginning to be made. However, the budget size is miniscule and no special organization for defense artificial intelligence (AI) exists. In addition, there is no plan to use defense AI in earnest in defense operations, and it is merely positioned as one of the technologies that attract widespread attention. Attempts to deepen the knowledge of artificial intelligence among SDF personnel have only just begun.

1 Thinking About Defense AI

A search of the database of the Nihon Keizai Shimbun (Nikkei), one of the most widely read newspapers by Japanese businesspeople, shows that the first article in which the words “artificial intelligence” and “defense” appear simultaneously was on 12 May 2014 (Nihon Keizai Shimbun 2014). After that, there are some reports on AI use on defense equipment and cyber defense.

Contrary to these public reports, aeronautical and space science journals discussed the defense AI much earlier. In 1983, Tamotsu Sakamoto of the third Research Institute of the National Defense Agency’s Technical Research Headquarters published a paper titled “On Navigation Using Artificial Intelligence Search Methods” (Sakamoto 1983). However, when this paper was written, deep learning was not yet in use.

A paper with a similar perspective was published in 1990. In their paper Akira Soga and Hideo Nakashima wrote, “artificial intelligence application systems, including expert systems, have been studied and developed overseas, especially in the United States, under the Strategic Computing Program (SCP) and other programs. In contrast, in Japan, research and development tends to be conducted mainly for industrial applications” (Soga and Nakashima 1990). The paper focused on systems for use on manned aircraft, and it does not examine the use of autonomous artificial intelligence to control current unmanned aircraft.

Japan’s defense policy is called “exclusively defensive defense.” The Japanese constitution after World War II does not allow Japan to have full-fledged military forces and the Japanese nation cling to peace-oriented diplomacy. The so-called Yoshida Doctrine, which was named after Prime Minister Shigeru Yoshida, was the basis of Japan’s diplomacy, defense policy and economic policy. The doctrine guided the Japanese government to allocate more resources towards the economy while keeping defense spending low under the nuclear umbrella based on the Japan-U.S. Security Treaty. That combination of policies enabled Japan to recover quickly from the destruction of World War II. This situation has been changing step by step as threat perceptions are altered by external incidents. However, there is a strong long shadow of the past in the use of technologies for defense purposes.

After the end of World War II, a general understanding emerged to avoid research on military-related technologies. Consequently, Japan has been reluctant to discuss military applications of new technologies, including AI. The Science Council of Japan was established in 1949, soon after the end of World War II. The following year, in 1950, the Council issued a “Statement of Determination to Never Follow Scientific Research for the Purpose of War,” and in 1967 it issued a “Statement of No Scientific Technology for Military Purposes.” This post-World War II atmosphere was based on the reflection that the academic community, including universities, had been mobilized for war.

However, with the outbreak of the Korean War in 1950, the General Headquarters of the Allied Forces (GHQ) decided to change the policy of demilitarization of Japan, and in August 1950, paramilitary organizations called the Japan Police Reserve Corps and the Coastal Safety Force were established by GHQ. The Police Reserve Corps was reorganized as the National Safety Forces in 1952 and reorganized as the Ground Self-Defense Force in 1954. At the same time, the Coastal Safety Force became the Maritime Self-Defense Force, and the Air Self-Defense Force was newly organized. The three Self-Defense Forces were not full military forces like those of other countries, but rather organizations based on the constitution with an exclusively defense purpose. It took 53 years for the Defense Agency, which oversaw the SDF, to become a ministry as the Ministry of Defense in 2007.

Defense spending, but not military spending, subsequently increased as Cold War tensions rose, and there were calls for more defense equipment. However, the response of the Science Council of Japan and universities was weak, and there was no atmosphere of active cooperation, even for the sake of exclusive defense.

When the Cold War ended with the Malta Summit, the fall of the Berlin Wall in 1989, and the dissolution of the Soviet Union in 1991, calls for an expansion of the defense industry weakened.

However, in 1998, when North Korea launched a Taepodong missile and it passed over Japanese territory, the North Korean military threat was rapidly recognized and came to be known as the “Taepodong Shock.” Simultaneously, China’s gradual military buildup began to be viewed as a cause for concern. By that time, voices denying the existence of the SDF had weakened in Japanese society; rather, the development of defense forces on an appropriate scale was considered essential. Nevertheless, there was a political consensus that defense spending should be limited to 1% of GDP, and this was maintained until recently.

However, when the second Shinzo Abe’s administration was formed in 2012, threats by North Korea and China became more obvious and the Abe administration set up “proactive pacifism” as a matter of defense policy. In December 2013, the Abe administration approved the National Defense Program Guidelines (NDPG). The document states:

From the aspect of security, it is necessary to utilize civilian technology effectively also in the field of security through regularly assessing the trend in science and technology including information related to technological development as well as consolidating the capabilities of the government, industry and academia. Under such recognition, the Ministry of Defense will strive to make effective use of civilian technology that can also be applied to defense (dual-use technologies), by enhancing partnerships with universities and research institutes, while strengthening technology control functions to prevent the outflow of advanced technologies (MoD 2013: 28).

However, the 2013 National Defense Program Guidelines contained no mention of artificial intelligence. Yet, while a shift towards defense technologies in government and academia is occurring, the debate over artificial intelligence gradually coincided. The first Defense Program Guidelines that mention AI were released in December 2018, also under the Abe administration, indicating that:

Due to advances in military technologies, a variety of threats can now easily penetrate national borders. States endeavor to develop weapons that leverage cutting-edge, potentially game-changing technologies. They also engage in research of autonomous unmanned weapon systems equipped with artificial intelligence (AI) (MoD 2018: 3–4).

It also stated that one of the measures to strengthen the basis of human resources is to “make focused investments through selection and concentration in important technologies including artificial intelligence and other potentially game-changing technologies” (MoD 2018: 24).

In response to the 2018 National Defense Program Guidelines, the Ministry of Defense showed plans to introduce artificial intelligence to use it to automate cyber defense, translate military and defense-related data, and manage equipment (determine repair points and the need to replace parts) (Nihon Keizai Shimbun 2019).

The National Defense Program Guidelines were renamed the National Defense Strategy when it was released 4 years later in December 2022. There is a reference to AI in the following statement:

Rapid advances in science and technology are fundamentally changing the paradigm of security. Countries are striving to develop cutting-edge technologies that could dramatically alter the character of warfare and thus prove to become ‘game changers.’ China in particular has been rapidly promoting accelerated technological innovation and its application for military purposes under the name of the ‘military-civilian integration strategy:’ China is notably accelerating military capability development premised on unmanned assets that leverage artificial intelligence (AI). These trends are resulting in fundamental changes to the way the military is organized as well as the way warfare is prosecuted (NSC of Japan 2022a: 5).

In the Defense Buildup Program announced with the 2022 National Defense Strategy, a more in-depth description of artificial intelligence can be found (NSC of Japan 2022b). First, as a response to information warfare, including in the cognitive domain, the Plan states,

In addition, the following functions will be developed: automatic collection and analysis of open-source information using artificial intelligence (AI), which will enable continuous collection and analysis of information on trends in each country; automatic collection of information on social networking sites, etc., to determine the authenticity of information communicated by each country; and future forecasting functions for estimating the security situation (NSC of Japan 2022b: 15).

In relation to command, control, and information-related functions, the plan seeks to “accelerate decision-making through the use of Artificial Intelligence (AI), etc., while strengthening the resiliency of the network.” In about 10 years, the report continues, Japan will “reinforce information gathering and analysis capabilities through the use of AI, etc., while enhancing the system for persistent information gathering and sharing” (NSC of Japan 2022b: 51).

Colonel Hiroshi Ito, who worked in cyber defense in the Japan Ground Self-Defense Force, notes that artificial intelligence will be used in all kinds of weapons in the future (Ito 2023: 68–70). AI will extend human intellectual capabilities as an aid to humans and will play the role of a calm responder on behalf of humans. He points out that at present AI has five advantages:

  1. 1.

    Can learn;

  2. 2.

    Able to handle a wide variety and large volume of data;

  3. 3.

    Has high processing speed;

  4. 4.

    Can share what it has learned with other AIs;

  5. 5.

    Has no human error.

The destination, he says, is the unmanned battlefield. However, there are concerns about autonomous weapons, and he appeals for an urgent debate on laws, ethics, and policies to shape the development of this technology.

The 2023 Defense White paper mentions AI nineteen times (including in the headlines). It is mentioned in the defense policy descriptions of AUKUS, Australia, China, India, the United States, and Russia. The section that mentions AI the most is Section 1 “Trends in Science and Technology Expanding to Information Warfare” in Chap. 4 “Trends in Space, Cyber, and Electromagnetic Fields, Information Warfare, and Other Issues for the International Community.” It covers about half a page and describes the defense applications of AI in various countries and internationally. However, there is little explanation of Japan’s own defense use of artificial intelligence. After reiterating the items presented in the Defense Buildup Program, the following statement is made in the “Efforts to Enhance Intelligence Analysis and Other Functions” section:

In order to win battles in a situation in which the battle situation will become more rapid and complex in the future, it is necessary to establish a system that enables real-time information sharing by making maximum use of various means, including artificial intelligence (AI), and further strengthening capabilities such as information gathering and analysis, and to continuously and more accurately grasp the intentions and capabilities of countries surrounding our country (MoD 2023d).

On 25 August 2023, the Japanese government held the first meeting of the Council of Ministers Concerned with Research and Development and Public Infrastructure Development that Contribute to Strengthening the Comprehensive Defense Systems (Cabinet Secretariat 2023). Among R&D for civilian use, nine areas, including AI and cyberattack countermeasures, were designated as “key technology issues” that will contribute to strengthening the defense system, and the government confirmed its policy to work on them across ministries and agencies (Table 1). At the meeting, the chairman, Chief Cabinet Secretary Hirokazu Matsuno, emphasized the importance of “breaking down the stove-piping between ministries and agencies and strengthening the comprehensive defense system in order to efficiently utilize the resources and capabilities of our entire nation” (Cabinet Secretariat 2023).

Table 1 Japan’s key technology issues
Full size table

2 Developing Defense AI

The Japanese government’s defense policies are presented in three defense documents: the National Security Strategy, the National Defense Strategy, and the Defense Buildup Program (MoD 2022b). The Ministry of Defense is responsible for their implementation, but the Acquisition, Technology & Logistics Agency (ATLA) is responsible for the technical aspects. ATLA’s Technology Strategy Department conducts the Short-term Demonstration Project for New Technology. It is a project to demonstrate the effectiveness of advanced technologies that are at the level of practical use in civilian applications by bringing together civilian engineers and operators in order to promptly solve problems faced by units and to promote their practical use in a short period of about 3 years. The project also aims to “reduce defense product prices and maintenance costs by using the results of this project in civilian markets, etc.” (ATLA 2023c). Table 2 provides an overview of the projects launched within this framework.

Table 2 AI Projects Launched by Acquisition, Technology & Logistics Agency
Full size table

ATLA began the Security Technology Research Promotion System in FY2015, which recruits and funds research and development through an open application process (ATLA 2023b). The launch of this system caused a major stir. This is because the Japanese academic community had been reluctant to research and develop technologies that could be diverted to military use, but the Ministry of Defense and ATLA decided to break that trend and provide its own R&D funds for technologies that could lead to defense capabilities. In FY2017, 109 applications were received, of which 58 (53%) were from universities and other institutions, 22 (20%) from public research institutions, and 29 (27%) from companies (ATLA 2015).

However, because applications to the Security Technology Research Promotion System stirred controversial media coverage, the number of applications in FY2016 decreased to 44, with 23 (52%) from universities, 11 (25%) from public research institutions, and 10 (23%) from companies (ATLA 2016). In FY2017, however, the numbers returned, with 104 applications (ATLA 2017).

The first research on AI under this program appeared in FY2018. Natsuki Matsunami of Mitsubishi Heavy Industries, Ltd, one of Japan’s largest defense contractors, was awarded the “Basic Research on Problem Coping by Collaboration of a Very Small Number of Humans and AI” (ATLA 2018). In FY2019, developing AI to understand human mental states was included in “Development of a Method for Estimating Latent Brain Dynamics and Elucidation and Control of Mental State Transitions,” for which Eiji Uchibe of Advanced Telecommunications Research Institute International (ATR) was awarded the contract (ATLA 2019b). Since 2018, 11 defense AI studies have been launched.

Despite a marginal number of applications, ATLA has shown strong interest in AI-related research. 31 research themes were listed by ATLA for the FY2023 open call, including the following 12 that directly refer to AI in the title or description (ATLA 2023a):

  • Fundamental research on sequential decision-making AI architectures that can build trust in unknown environments

  • Basic research on AI to realize accurate predictions from all kinds of information

  • Fundamental research on AI architecture with robustness in unknown environments

  • Basic research on the improvement of cognitive and communicative functions through brain science

  • Basic research on human cognitive support for multiple unmanned aircraft operation and control

  • Basic research on cognitive security

  • Basic research on security that automates protection against unknown attacks on wireless communications and cyber kill chain fragmentation

  • Basic research on magnetic sensor technology

  • Fundamental research on technologies for understanding materials and objects in the ground or on the seafloor

  • Basic research to significantly improve the performance of unmanned underwater and surface vessels and underwater vehicles

  • Basic research to significantly improve the performance of aircraft and unmanned aerial vehicles

  • Basic research to significantly improve the performance of vehicles and unmanned vehicles

It is important to note that none of these calls and the respective projects are directly related to weapons. With a sense of aversion to military-related research still lingering, researchers remain reluctant to directly engage in military technology research. Against this backdrop, open calls are being made for technologies that are less military in nature but still applicable to defense.

In addition to this funding for external research, development is also taking place within ATLA. At the Agency’s Technical Symposium 2019, AI was mentioned in three of the 16 oral sessions and 23 poster session summaries, thereby covering, among other things, the following topics:

  • Taisuke Katayama, “R&D Vision: Achieving Multidimensional Integrated Defense Capability and Beyond”

  • Masataka Okubo, “An Attempt to Introduce Artificial Intelligence (AI) into Ship Design (Demagnetizer Gear Design)”

  • Toshiro Kamitani, Hiroka Sano, and Kenji Hamano, “Target Categorization from Synthetic Aperture Radar Images”

According to a report of the second topic, the work that had been done by skilled technicians was replaced by AI, and tens of millions of verifications can now be done in three to four hours, a significant time saving. A report of the third topic examined the possibility of using AI to classify radar images (ATLA 2019a). These reports are also being used within ATLA to meet actual defense needs.

Japanese defense companies are working on AI, although not necessarily for defense equipment. The demand for Japanese defense companies comes primarily from the civilian sector. AI is being developed for such civilian demand, and there is a high possibility that such technology and knowledge will be applied to defense equipment if necessary.

Moreover, Japan also looks at international cooperation to develop emerging technologies and AI. In August 2023, Craig Martell, Chief Digital and AI Officer of the Pentagon, visited Japan, Singapore, and South Korea to discuss opportunities to “deepen cooperation associated with data, analytics, and the responsible deployment of AI” (Vincent 2023). These talks mirror long-standing bilateral interests in jointly advancing dual-use technology cooperation (Tajima 2023). To this purpose both countries signed a “project arrangement” in late December 2023 to jointly develop AI for the use of Unmanned Aerial Vehicles (UAV) “that will be used in a ‘loyal wingman’ role alongside” the Global Combat Air Program (GCAP) (Kadidal and Kumar 2023). US-Japanese defense AI cooperation could become even more relevant should Japan join trilateral cooperation with Australia and the UK in the AUKUS framework.

In parallel to technology cooperation with the US, Tokyo and London have been extending technology ties as well. In May 2023, both nations adopted the “Hiroshima Accord,” which outlines a strategic technology partnership, and renewed a science and technology agreement to cooperate on innovation and new technologies (Prime Minister’s Office 2023; Evenstad 2023). Expanding science and technology bonds are relevant given trilateral cooperation with Italy on the GCAP (Chuter 2022) and recent developments in Japan’s strategic environment. In view of both aspects General Jim Hockenhull, Commander of the UK Strategic Command, recently underlined that by “using both the Japanese and the UK industrial base, (both nations) can generate even greater and even better capabilities, which play a part in any deterrence approach” (Kitado 2023).

In addition, Japan also engages with France on AI cooperation. According to ATLA, bilateral research cooperation includes mine-countermeasure technologies, with AI being used to “identify targets from images obtained by mine detectors” (Majumdar 2023: 25).

3 Organizing Defense AI

In 2016, at the direction of Prime Minister Shinzo Abe, the Japanese government created the Strategic Council for AI Technology, which brings together the wisdom of industry, academia, and government and eliminates organizational silos (Cabinet Office 2018). On 31 March 2017, the Council released the Artificial Intelligence Technology Strategy presenting a three-phased approach: First, data-driven AI utilization will progress in each domain by around 2020; second, general use of AI and data will progress beyond the boundaries of individual domains by around 2025; and third, each domain will be connected in a complex manner by around 2030, creating an ecosystem (SCAIT 2017).

Two years later, the Council released the “Strategic Action Plan for Artificial Intelligence Technology” (SCAIT 2018). The plan aimed at promoting AI involving a diverse set of ministries—Ministry of Internal Affairs and Communications, the Ministry of Education, Culture, Sports, Science and Technology, the Ministry of Economy, Trade and Industry, the Cabinet Office, the Ministry of Health, Labor and Welfare, the Ministry of Agriculture, Forestry and Fisheries, and the Ministry of Land, Infrastructure, Transport and Tourism – the action plan primarily targeted the industrial use of AI whereas AI for defense was not considered.

Based on this strategy, AI research was to be conducted at three national research institutes:

  • First, in April 2017 the National Institute of Information and Communications Technology (NICT) established the AI Science Research and Development Promotion Center (Kidawara 2019). However, NICT is a research institute under the Ministry of Internal Affairs and Communications (MIC) and does not conduct defense-related research, apart from cybersecurity.

  • Second is the Artificial Intelligence Research Center (AIRC) of the National Institute of Advanced Industrial Science and Technology (AIST). The center studies the application of artificial intelligence to the manufacturing, service, medical and nursing care, and security sectors. However, there does not appear to be any research focused on defense or military applications. The applications mentioned for the security sector are automatic explanation functions for videos and evacuation guidance during disasters.

  • Third is the Center for Advanced Intelligence Project (AIP) of the National Institute of Physical and Chemical Research (RIKEN). The Center has three research groups: the Generic Technology Research Group, the Goal-Oriented Technology Research Group, and the Artificial Intelligence in Society Research Group. However, these groups do not appear to be conducting research directly related to defense.

On 11 June 2019, the Japanese government, via the Integrated Innovation Strategy Promotion Council, released the AI Strategy 2019: AI for People, Industry, Region, and Government (IISPC 2019). The document focuses on education and industry; again defense applications were not considered. Two years later, in June 2021, the Japanese government published the AI Strategy 2021, but again, the Ministry of Defense’s efforts were not explicitly mentioned. In response to this AI Strategy 2021, the AI Strategy Executive Council established the “New AI Strategy Study Council”. The AI Strategy 2022 was the first capstone document to refer to AI in relation to security with the following statement:

In view of the increasingly complex international situation and changes in socio- economic structures, various initiatives are being considered for important technologies including AI from the perspective of economic security. Therefore, it is necessary to coordinate related measures so that the government as a whole can effectively prioritize them (IISPC 2022a).

However, this refers to economic security (or geo-economics), not military security. The following statement was also included.

Pursue fusion of AI and Japan’s strengths to address challenges unique to Japan ([1] health, medical care, and nursing care; [2] agriculture; [3] infrastructure and disaster prevention; [4] transportation infrastructure and logistics; [5] regional revitalization; [6] manufacturing; and [7] security) (IISPC 2022a: 27).

Attached to the 2022 strategy was a “List of AI Strategy 2022 Initiatives,” in which the Ministry of Defense included a section on “Promoting Research on the Application of AI Technology to Contribute to the Defense of Japan” (IISPC 2022b). This is the first documents that brings AI and defense together and underlines the Japanese government’s growing awareness of addressing this issue.

According to the AI Strategy 2022 progress report published in April 2023, the Ministry of Defense is addressing defense AI by bringing in experts from the outside and stepping up training efforts. In addition, the Ministry also intends to enhance the capabilities of equipment used by the SDF, including detection and identification (IISPC 2022c).

On 11 May 2023, the Japanese government held the first meeting of the AI Strategy Council of experts together with the relevant Ministers. Prime Minister Fumio Kishida attended the first meeting. About 2 weeks later, the Council met again, discussed several topics and issues, among others, the following statements:

AI is also an important tool in global issues such as security, disaster management, and global warming countermeasures, and our country needs to work on technological innovation together with volunteer countries (AI Strategic Council 2023: 6).

There is an argument that the use of AI is important in security-related matters as well, but it should be handled flexibly, for example by leaving it for discussion by specialized departments, depending on the need for information management (AI Strategic Council 2023: 16).

In general, Japan’s AI focus rests with industrial applications and educational efforts. Research to advance defense AI has only just started in 2022. Although no dedicated institutions addressing AI within the defense establishment have yet been created, the 2022 National Security strategy indicates the willingness of the Ministry of Defense to “establish a mechanism to aggregate military information” and urges the intelligence and policymaking departments to “enhance information management analysis” (Majumdar 2023: 25).

4 Funding Defense AI

No summarized data has been released on the Japanese government’s AI budget. According to the Nihon Keizai Shimbun (Nihon Keizai Shimbun 2023), the Japanese government has summarized its basic policy and budget for policies related to generative AI for the FY2024 budget estimates. The government will focus on the development of infrastructure for the development of AI, including supercomputers and high-quality data, and will create a foundation for research and development in Japan to reduce its dependence on foreign countries for AI development.

The government’s overall AI-related budget was approximately JPY100bn (or €630M) in the initial budget for FY2023. The government is considering doubling the budget to JPY200bn (€1.26bn) in FY2024. The basic policy has three pillars to strengthen the research and development infrastructure for generative AI by developing of data centers with supercomputers, developing high-quality AI training data, and creating generative AI that can be used in scientific research.

In March 2023, the Ministry of Defense included a section on defense AI in its budget (MoD 2023b, c). The overall amount spent on defense AI is JPY6.58bn (€41.4M), or around 0.096% of the total FY2023 defense budget of JPY6.8219tn (€42.9M). The defense AI budget encompasses the following items:

  • Development of functions for automatic collection and analysis of public information using AI: JPY2.2bn (€13.9M)

  • Research on speeding up decision making using AI: JPY4.3bn (€27M)

  • Utilization of external forces for AI implementation: JPY50M (€0.32M)

  • Training of AI human resources by providing AI training courses: JPY30M (€0.19M)

In addition to this, part of the “Project for Establishment of Systems for Strengthening Production Base of Defense Equipment, etc.” (JPY36.3bn or €226M) will be used to “implement efficiency improvement of defense equipment manufacturing process, etc. through introduction of advanced technologies such as 3D printer technology and AI technology.” Thus, the exact size of the defense AI budget is not publicly known as spending on AI in other budgets, for example on cybersecurity, need to be considered as well.

5 Fielding and Operating Defense AI

In April 2022, the Ministry of Defense held a presentation entitled “AI Initiatives in the Ministry of Defense” (MoD 2023a). The document states, that:

In the field of national defense, AI technology is expected to be utilized for faster and more efficient information processing, situational assessment, operational planning, and high-level search and rescue using unmanned aircraft, etc. Many countries including China and the United States are actively investing in R&D related to AI technology as it could change the future of warfare.

The Ministry of Defense also believes that AI technology can be a game changer and that it is necessary to make focused investments and realize its implementation in defense applications as soon as possible. In this regard, the document also illustrates two practical examples illustrating exemplary lines of effort that the SDF are interested in:

  • The first example refers to research on AI-based radar image identification to study the technology to automate the identification of radar images using AI to efficiently conduct constant and continuous information gathering and warning monitoring activities. By automating the deciphering and identification of radar images, which requires skill, it is expected to reduce the burden on units and improve mission efficiency.

  • The second example refers to a study of the components of an unmanned submersible monitoring machine. This will involve research on autonomous monitoring technology and sensor systems. The plan is to apply AI technology to decisions about the behavior of underwater vehicles used for long term surveillance.

In addition to these two examples, the Ministry’s presentation also discusses additional initiatives summarized in Table 3.

Table 3 AI efforts of the Japanese Ministry of Defense
Full size table

6 Training for Defense AI

We were unable to find any direct documentation of how the Ministry of Defense and the SDF are training for the advent of defense AI. However, the SDF does recruit publicly for senior positions and mentions AI in current job openings as the examples illustrated in Table 4 and published in August 2023 highlight.

Table 4 Job openings related to AI at Japan’s Self-Defense Forces
Full size table

A public notice issued by the Ministry of Defense in December 2022 indicated a competitive bid to implement basic training for AI related to human resources development (MoD 2022a). This initiative exemplifies the Ministry of Defense’s effort to recruit instructors from the private sector to help train the Ministry of Defense and the SDF. In addition, the Department of Information Engineering at the National Defense Academy, which trains many of the Self-Defense Forces officers, offers AI as an elective compulsory subject (National Defense Academy 2024).

At a more general level, Air SDF LTC Kenshi Kamitakahara, who belongs to the Operational Theory Laboratory of the Aviation Research Center, a research institute within the Air SDF, recently touched upon another aspect related to training: the lack of data. In view of using machine learning related to defense equipment he underlined the need to “collect as much data as possible and extract high quality data from it” (Kamitakahara 2021a). In another paper, he focused on AI and defense simulations and argued that “AI for defense simulations is considered to be the most technically challenging area, as decisions to be made based on the simulation results are extremely critical” (Kamitakahara 2021b).

7 Conclusion

Although Japan is often recognized as one of the most technologically advanced countries, there is still strong hesitance to the use of technology for defense purposes. Universities and other research institutions are not necessarily active in R&D for defense uses.

On the other hand, interest in AI is growing, as in other countries. However, this interest is skewed toward educational and industrial applications. The defense industry is working on AI across the board, but Japan’s defense industry has a large proportion of civilian demand, and there are aspects of the industry that are not necessarily for defense use. If necessary, they will be applied to defense applications, but they are not actively promoted as AI for defense.

The Ministry of Defense is gradually beginning to show interest in AI, as highlighted by a growing number of public documents referring to AI published since December 2022. Although still small, dedicated defense spending on AI is ticking up.

In general, Japan is a latecomer on the growing international defense AI scene—with no major achievements yet to see. However, the security environment surrounding Japan is worsening, and China’s military spending is increasing remarkably. China has also pointed out that it will use and incorporate AI for military purposes. Thus, for Japan to respond to what is going on in its strategic neighborhood, next-generation technologies are indispensable to expand its defense capabilities. In this regard, AI and other technologies are of strategic importance.

In so doing, ATLA is Japan’s core organization to develop defense AI. ATLA should promote defense AI initiatives that reflect the strategic thinking of the Prime Minister’s Office and the Ministry of Defense. In response, the academic community, including national and private universities, and the private sector should deepen their cooperation. Japan’s unique post-World War II political environment should not be ignored. As the path toward war should be avoided, stepping up research and development efforts to the benefit of new technologies is indispensable to deter current and future threats.