By Tony Tai-Ting Liu

In the past century, concerning the issue of war and peace, technology has played a major role in implicating the outcome on the battlefield. For example, development of the U-boat submarine provided Germany with a formidable weapon in the high seas in the first and second Great Wars. Britain’s production of the modern aircraft carrier – the HMS Hermes – in 1917 opened up another path for competition in the future. The vehicle later evolved into a key indicator of hard power over the past decades. Introduction of the atomic bomb in 1945 in the cities of Hiroshima and Nagasaki demonstrated not only the superiority of United States’ military prowess at the time, but also the game changing quality of the bomb in future conflicts. Drones and stealth aircrafts are the latest innovations in the long line of new technologies that may change the nature of warfare in the near future.

Currently, much attention centers on the potential of artificial intelligence (AI) to become the next game changer in international relations among other new technologies. Founded on machine learning and big data, AI is considered by many observers as having the ability to transform international relations in at least three fundamental ways. First, in terms of decision making, particularly during times of crisis, compared with humans, AI has the potential to comprehend and digest a large amount of data without being affected by emotions or other exogenous factors. Decision making in realms such as foreign policy or combat can potentially be carried out more efficiently and effectively, thus greatly increasing the speed of developments in international relations and reducing the reaction time of states. Second, in terms of national security, integrated with technologies such as surveillance and facial recognition, AI may contribute to fields such as anti-terrorism and border control. If integrated with drone technology, AI may become an effective tool in combating terrorism and criminal activities.

Third, and perhaps most importantly, from a geopolitical perspective, AI is in the midst of triggering the next round of strategic competition in international relations in the near future. Since the release of respective national AI strategy by Canada and Japan in March 2017, almost three dozen countries followed suit and announced their national AI strategies. A number of states and significant international actors have joined in the AI game since 2017, including the US, China, the European Union, the United Kingdom (UK), Germany, France, Sweden, Finland, India and Australia. Premised on the political, economic and social implications AI may have on states, it seems clear that governments are becoming increasingly aware of the potential of AI technology to generate a new round of power shifts in the world. Despite a strong emphasis on military power in the realm of geopolitics, in the new century, power has expanded beyond weapons and firepower to encompass economic, cultural and media influences that have the ability to implicate geopolitics.

Like many technologies that came before, AI holds the potential to change the world immensely. Yet in contrast with nuclear, internet, digital and telecommunication technologies that fueled the latest rounds of social revolution, the impact of AI may be even greater, in the sense that the technology touches on the issue of intelligence, or a quality that is traditionally considered to be reserved for the sentient human being. As a highly intelligent species, the human being is capable of harnessing technologies to improve his wellbeing. Hence as a technology that can compete and potentially overtake, or essentially become smarter than human intellect, AI has generated much debate across the world. Meanwhile, observers are already imagining the application of AI to everyday life in the future. The potential harnessing of AI in various fields is where the transformative power of the technology lies.

As revealed in AI: Into the New World, a documentary released by Japan’s NHK broadcasting group in 2018, since the showdown between professional Go (圍棋) master Amahiko Sato and AI, scientists and engineers have put the technology to test in various functions. For example, integrated with robotic technology, AI can be seen in application through humanoid robots such as Pepper and Sophia. The ability to talk and interact separate humanoid robots from traditional machines and make them potential “beings” in future society that can serve in various capacities. On the other hand, combined with big data, AI has the ability to guide taxi drivers to routes that have relatively higher possibility for intercepting potential passengers. In such sense, the ability of AI to guide drivers more “scientifically” makes the technology more efficient than simple human experiences.

In terms of urban planning and so called “smart cities” in the future, AI is a key technology in current imaginations. A large part of present imaginations is on autonomous vehicles or self-driving cars that will be cleaner and more efficient. Smart autonomous vehicles could greatly reduce traffic jams and accidents that implicate the city. In addition, equipped with AI sensors that can track the health condition of passengers, should one feel unwell aboard the autonomous vehicle, immediate aid and relief can be provided. By adopting autonomous vehicles in the future, the general hope is that human errors can be greatly reduced, which could in turn lead to increase public safety and more human energy to focus on other tasks. Transportation is one of many fields that has high potential of undergoing transformation with the adaptation of AI.

Noting the potential impact AI has on society, the development of AI technology and its practical application suggests enormous business opportunities waiting to be excavated. Hence from an economic standpoint, perhaps reminiscent of the dot com boom and the rise of digital technologies, many states have seen the potential of AI and have identified the technology as an important part of their modernization and development plans. A clear example is Made in China (MIC) 2025, a national strategic plan aimed at upgrading China’s manufacturing sector proposed by Beijing in 2015. MIC 2025 identifies information technology and robotics – industries that involve AI – as priority industries to be developed in the near future. In 2016, following proposals such as the New Robot Strategy (2015) and Industry 4.1J, the Japan Council for Science, Technology and Innovation proposed Society 5.0, an initiative that seeks to establish “a human-centered society that balances economic advancement with the resolution of social problems.” AI is identified as a critical component for achieving the balance.

Perhaps with the advancement in robotics, AI suddenly re-emerged as a heated topic for discussion. It should be noted that the first serious discussions centered on AI came about more than two decades ago, highlighted by Stuart Russell and Peter Norvig’s Artificial Intelligence: A Modern Approach.[1] The fact that the technology was decades away from maturation and the world indulged more in the digital revolution may have possibly delayed general attention on the military potential of AI. In some sense, the prioritization of digital technologies may also be a choice of convenience, as digitalization could complement the development of AI by pooling ideas and establishing the platform that AI could work on.

Regardless of the reasons for the coming of AI, a major factor for the securitization of AI or making AI a national security concern, is the China factor. Due to the debate on China’s rise in the new century, China’s every move is scrutinized, with AI being the latest development under watch. While there is no ground to discriminate China on its potential use of AI, observers suspicious of the Chinese Communist Party (CCP) and its ulterior motive of revisionism point to examples of AI being a particular lethal weapon for Beijing. For example, Mozur points out the combination of AI and facial recognition technology to target the Uyghur minority in China, a group that has suffered immensely under the Xi Jinping regime. On the other hand, Abadicio points out the military application of AI to unmanned combat aerial vehicles (UCAV) or combat drones, a weapon that is seeing increasing action on the battlefield. In 2019, the US-China Economic and Security Review Commission - a bipartisan group established as part of the 2001 US National Defense Authorization Act - warned of China’s growing capabilities in AI in its annual report, noting that “China firms and research institutes are advancing uses of AI that could undermine US economic leadership and provide an asymmetrical advantage in warfare… [and] rapidly modernize its military.”[4]

In July 2017, following in the footsteps of Canada, Japan and Singapore, China became the fourth country in recent memory to release a national strategy for AI. Despite not being the first country in the world to announce an AI strategy, China seems quite active nonetheless, announcing the Next Generation AI Plan in July and a corresponding Three-Year Action Plan to propel the development of AI in December. According to the OECD AI Policy Observatory, as of February 2020, more than 60 countries, including Taiwan, have initiated over 300 AI policy initiatives covering various sectors such as agriculture, education, environment, health, trade and transport.[5] The OECD notes the US as the most active state, proposing 40 AI related initiatives. The United Kingdom, Germany, Singapore, Australia and Colombia have all announced more than a dozen initiatives concerning future development centered on AI, attesting to the emphasis that the global community places on AI. While not all initiatives are geared towards military advancement, the fact that more than a quarter of all countries in the world have their respective AI initiative, including all major states with competitive capability for research and development (R&D), demonstrates the current race for AI.

In terms of R&D concerning the potential military application of AI, the US seems to boast the most comprehensive set of initiatives among all competing states.

On February 11, 2019, US President Donald Trump signed Executive Order 13859 announcing the American AI Initiative. The Initiative directs the federal government to pursue five pillars for advancing AI: (1) invest in AI R&D, (2) unleash AI resources, (3) remove barriers to AI innovation, (4) train an AI-ready workforce, and (5) promote an international environment that is supportive of American AI innovation and its responsible use.[6] Among the 40 initiatives identified by the OECD, at least 5 initiatives are directly related with the advancement in national security, including the Defense Advanced Research Projects Agency’s (DARPA) open call for AI related research projects, the Department of Defense’s (DoD) establishment of an AI strategy and a Joint AI Center (JAIC), the American AI Initiative (also known as the Executive Order on Maintaining American Leadership in AI), and establishment of the National Security Council on AI (NSCAI). Currently, the US is set to increase AI spending by 34% for fiscal year 2021, with the DoD’s budget increase from 780 million USD in 2020 to 841 million USD in 2021, JAIC’s budget increase from 242 million USD to 290 million USD, and DARPA’s R&D investment in AI to increase from 459 million USD to 500 million USD.[7]

While one of the most popular buzzwords in strategic studies in recent years may be AI, it is important to critically reflect on history and an arms race driven by advancements in technology. Development of the Strategic Defense Initiative (SDI) by the Ronald Reagan administration in the 1980s is worthwhile noting. Also known as the “Star Wars Program,” the initiative was inspired by the popular sci-fi movie series Star Wars, in which the plot is based in a distant land and involved combat in outer space. With the evolution of the nuclear arms race between the US and the Soviet Union into a space race at the time, the overarching idea was that the US could defend against a nuclear missile attack by sending advanced devices into outer space, which would have the ability of detecting and countering incoming attacks. Just like the movie, the space device will annihilate incoming missiles in mid-air through the use of laser beam.

The SDI was conceived by the Reagan administration as a game changer that would effectively set up the US in a good position to launch a second strike against the enemy. In a broad sense, introduction of the SDI initiated a new round of war gaming between the US and Soviet Union, grounded on the belief that the SDI was the answer for safeguarding against a nuclear strike. Such belief extended to scenarios for extended nuclear war in which second and third strikes were possible, which in turn further reinforced the logic of mutual assured destruction or MAD. Announcement of the SDI took the nuclear arms race into another dimension, which perhaps served as the distant reason for the current space race among states. Nonetheless, perhaps a more important implication of the expanded arms race was the eventual bankruptcy and collapse of the Soviet Union, as competition with the US in the new field was simply economically unsustainable.

Considering the current race towards AI in light of the fleeting episode of the SDI race in the 1980s, the dominant implication seems obvious – is AI really the key to the future, or is it merely another dead end played up by states that are driven by other interests? Comparison to the SDI should not be dismissed easily, as the current development and discourse over AI exudes striking semblance with the past. For example, while AI may hold much untapped potential, maturation of the technology for military application seems to be better realized in words rather than actual progress at the moment. After all, merely three years have passed since the matchup between Amahiko Sato and AI, and to a large extent, states are still exploring the effectiveness of drones and other new technologies on the battlefield. A clear gap remains, as there has yet to be a conflict that involves only self-propelling intelligent machines thus far, or even a conflict that partly involves intelligent machines. Regardless of developments, forward thinking corporate leaders such as Tesla’s Elon Musk and Alphabet’s Mustafa Suleyman, along with leading scientists and engineers, are already calling for the United Nations to ban the use of lethal autonomous weapons, or simply “killer robots,” to prevent third age war.[8] Such concern is supported by Human Rights Watch, which released a 55 page report in August 2020 that intends to raise general awareness and advocates a ban on killer robots.[9]

In essence, by relating the development of AI with the narrative of national security, or securitizing AI, the result could only be the establishment of a security dilemma as states race to acquire the technology, noting it as the future. Such dilemma may be acceptable if its effects are limited to economics, entailing a resource competition, or competition among states on the amount of investment that can be devoted to R&D on AI. The SDI suggests that a prolonged resource war could effectively collapse a superpower. Nonetheless, perhaps a more severe challenge comes from the growing acceptance that military imaginations and experiments can be carried out with AI. For example, if the killer robot is too dangerous and unethical, would the anti-ballistic missile system that is strengthened by AI be more acceptable and less controversial? What about the AI powered drone that has the ability to carry out limited or precision attacks? In a sense, weapons of mass destruction (WMDs) are unacceptable today due to their large-scale destruction of humans. However, the current fervor over AI has prompted many to pass over ethical issues such as the fusing of AI and existent weapons that may be defensive and less lethal.

On the other hand, the discussion of perception in relation with AI, interestingly, remains limited at the moment, despite the implications of perception for the making of foreign policy. As suggested by Robert Jervis, the entire Cold War could be explained by the perception and misperception of the superpowers.[10] In other words, state leaders perceived the world differently and such difference could serve as the driver of conflict. Concerning AI, the important questions are how the technology perceives and whether the human individual is willing to accept the perception of AI. Moreover, as politics necessarily entail decisions based on the prioritization of values, it remains unknown whether AI has the ability to distinguish and prioritize an assortment of values, which would implicate the final decision. Such consideration is not merely emotional, but asks the question of whether efficiency – a value that AI excels in – should be prioritized under all circumstances, even in the case of war.

On January 18, 2018, the Executive Yuan introduced a four-year (2018-2021) AI Action Plan aimed at boosting the competitiveness of Taiwan’s industries. In addition to transforming industries through the introduction of AI, the Action Plan also aims to promote Taiwan’s leading role in AI by increasing the number of related talent in universities and research institutions, building the island into an AI innovation hub, and liberalizing relevant laws and opening test sites.[11] In October, the Ministry of Science and Technology (MOST), responsible for the nation’s R&D, announced the formulation of a five year 517.5 billion USD “grand strategy for small country” centered on realizing the AI Action Plan. The strategy seeks to establish a high-speed computing platform for R&D, four AI innovation research centers and an AI robot makerspace among other plans.[12] It is clear that in terms of government thinking and state policy at least, Taiwan is very much in the global race for AI.

Beyond thoughts on the civilian use of AI, however, discussions on the military application and security implications of AI related to Taiwan remains relatively limited. The Institute for National Defense and Security Research (INDSR) noted the development of AI and its security implications in its annual evaluation report on trends in national defense technology. In September 2019, the National Chung-Shan Institute of Science and Technology, the nation’s top research institute, announced the “smart national defense” plan, a ten-year program aimed at the integration of technologies in response to the rapid transformation of warfare in the new age. In the local academia, discussions remain far and few.

Nonetheless, if China is considered as a main reason for the race for AI and a proactive actor in military innovations in relation to AI, the challenge for Taiwan is real and immediate. A predominant reason is Beijing’s continued threat for reunification by force, an option that has never been relinquished by Chinese leaders. Noting the use of AI technology in surveillance against the rebellious Uyghur minority, one wonders whether Taiwan may become a future testing ground for Chinese technology. For example, an easy way for China to influence Taiwan is to initiate an information war in cyberspace. AI can be used to generate discourse through social media that has the power to implicate general discourse in Taiwan, particularly noting the fact that Taiwan is a democratic society that is susceptible to public opinions. Another way is China’s potential use of AI drones in future conflicts with Taiwan. As recent as October 2020, the People’s Liberation Army Air Force (PLAAF) is reported to provoke Taiwan by flying near or into the latter’s air space. It is not unimaginable that drones and other smart weapons may be used against Taiwan in the future.

Despite the potential challenge that China poses, it is also true that AI also holds opportunities for Taiwan’s national security. Considering the fact that the military balance has tipped in favor of China over the past decade, advancements in AI may reinforce Taiwan’s efforts towards re-establishing an asymmetric balance with China. As new age war is not entirely traditional, if Taiwan can achieve breakthroughs in R&D on AI, the island may have a chance at countering China’s information war. On the other hand, drones may be another field that Taiwan has great potential in making advancements in. Noting Taiwan’s good relationship with the US, Japan and a number of European countries, collaboration with partner states in the joint advancement in AI application may be possible in the near future.

Assistant Professor, Center for General Education, National Chung Hsing University.