Media  Foreign Affairs and National Security  2019.01.23

How will AI change international politics?

the japan times on January 16, 2019

Another artificial intelligence-related exhibition will be held in Osaka next week. The second Kansai AI and Business Automation Fair features such AI-assisted services as robotic process automation-based labor saving and developing industrial robots with "deep learning" technologies. Here they go again.


This is the third AI boom in Japan. Unlike the first two booms in the 1950s and '80s, however, this time seems to be real. Recent technological breakthrough in image recognition, deep learning and big-data processing capacity is making this happen. Unfortunately, in the case of Japan it only happens in the private sector.


As a political scientist specialized in national security strategies, I am not satisfied with such AI applications in Japan. While enormous amount of human and financial resources has been invested in other major nations on the studies of military applications of AI, Japanese AI experts, to my surprise, don't seem to be interested.


More importantly, AI technologies will not only change military tactics but also alter national security strategies in the long run. This is the reason why last year I wrote a book titled, "AI and Neo-Geopolitics." AI is not a simple military means or tactical weapon. It could be a game changer in international politics in the future.


How will the AI revolution, as I asked in the book, change global power politics? Let's take the example of U.S.-China relations. From a geopolitical point of view, the history of China was that of competition among agricultural and nomadic land powers, and it is only recently that China became exposed to challenges by sea powers.


China lost Hong Kong to the British in 1842. Eleven years later, Commodore Matthew Perry's American black ships came to what today is Yokosuka, Kanagawa Prefecture. Until the 1990s, however, the United States and China were not geopolitical rivals. A great land power, pundits believe, can never be a great sea power simultaneously.


The exception was the United States, we thought, but we may be wrong. The reason why the faraway Western powers could challenge the Chinese land power was because they had achieved navigational technology breakthroughs. By the same token, China now could beat U.S. hegemony in the Western Pacific with its AI/big data technologies.


The low-tech land power of China is now catching up with and may even excel the high-tech sea power of the U.S. by overcoming the geopolitical vulnerability that has existed since the 19th century. This, however, may not apply to the case of Russia, which is technologically falling far behind China.


The AI revolution may also change the regional geopolitical power balance. For example, in East Asia, AI could drastically alter Japan-China relations. AI-related technological revolution has been so remarkable that China might overcome the regional geopolitical disadvantage vis-a-vis Japan in the years to come.


The most concerning element in the Chinese AI application is the breakthrough in the Chinese Communist Party's capability to monitor, identify, locate and finally arrest whoever is deemed hostile by the regime. Free of the need to protect its 1.4 billion citizens' privacy, Beijing can easily reinforce its authoritarian social control system.


After completing its public security system, China's next target will be South Korea and, of course, Japan.


Although historically Japan and China did not pose a serious threat to each other, China has become a present and existential threat to Tokyo since 2010. The Chinese AI revolution could further deteriorate the already grave situation.


If so, what should Japan do? As I wrote in my book, I found a hint from a paper by Harvard Kennedy School's Belfer Center for Science and International Affairs titled "Artificial Intelligence and National Security." The following are the lessons the authors learned:


1. Radical technological change begets radical government policy ideas.


2. Arms races are sometimes unavoidable, but they can be managed.


3. Government must both promote and restrain commercial activity.


4. Government must formalize goals for the safety of technology and provide the means to achieve these goals.


5. As technology changes, so does the U.S.'s national interests.


The paper made recommendations, such as that the Department of Defense "should fund diverse, long-term-focused strategic analyses on AI technology and its implications," and the "U.S. defense and intelligence community should invest heavily in "counter-AI"capabilities for both offense and defense." Oh, my! These are not applicable to Japan.


The conclusive connotations in my book's first chapter are as follows:


If AI changes national security strategies, it means that AI weapons could replace nuclear weapons in the future as a "strategic weapon" that can eliminate the enemy's will to fight.


If AI weapons eliminate the enemy's will to fight, it means that a nation's AI weapons can achieve the enemy's "mass destruction" without using nuclear weapons.


While nuclear weapons are becoming more and more unusable, AI weapons could achieve -- more easily and only with the judgment of the country wielding them -- the enemy's "mass destruction," which has long been considered taboo. For a nation to prevent this from happening, it must be able to deploy its own counter-AI weapons to offset the enemy's AI military capability.


For the past several decades, the global balance of power has been maintained with nuclear deterrence among the nuclear powers or the concept of mutually assured destruction. AI military applications may fundamentally change this balance of terror.


If a small nation or non-government entities acquires AI strategic weapons -- something that I cannot imagine taking place now -- what would happen? What if the AI weapons go out of control and activate themselves? Will AI deterrence power really work in the real world? Questions go on and on.


These are the ideas I have on artificial intelligence from a geopolitical and strategic perspective. Next week, I will focus on the socio-political aspect of AI applications. How do they affect the life and society of people? Are AI technologies a friend of mankind or an enemy ? Stay tuned.