Accelerating AI battle between the US and China... Tower of Babel Desire and Instinct of Distrust

Instincts of desire and distrust more powerful than the lessons of the Tower of Babel... The AI battle in the U.S. is actually accelerating.

(Seoul = Yonhap News) Senior Reporter Seungwoo Lee = As artificial intelligence (AI) is developing at a frighteningly fast pace, the speed control theory that AI development must be slowed down has emerged. As the rate of entrusting even AI development to AI increases, there is even an extreme prediction that one day an 'intelligence explosion' through recursive self-improvement (RSI) will occur and humanity will be destroyed or alienated. The sense of crisis has grown to the point where not only academia but also the industry is demanding that AI development be controlled. Recently, 22 AI scholars and big tech executives jointly published a paper warning of the risks of reckless AI development and urging the establishment of safety measures.

In addition to scholars such as 'AI father' Professor Emeritus Geoffrey Hinton of the University of Toronto, working AI technology leaders such as Antropic co-founder Jack Clark, OpenAI chief scientist Jakub Pahotsky, Microsoft (MS) Chief Science Officer (CSO) Eric Horvitz, and Vice President of AI Research Dawn Song Meta participated in the paper. Even though these are people who have led the creation and development of AI internationally, they are claiming that “now is the time to take a break.” Their concerns can be summarized as the fact that it is difficult to supervise the actions of AI agents with human capabilities, reaching a crisis where other AIs must be deployed to monitor them.

In this way, AI has emerged as a tangible threat to humanity. However, whether the speed control theory actually works in reality is another question. Many predict that speed control will ultimately be difficult due to human nature being difficult to change. Even if there is a justification and the necessity is recognized, the DNA of human nature is ultimately designed to pursue utility. No country or company wants to miss out on the hegemony of AI technology, which is clearly predicted to dominate all fields in the future. This is because if you fall behind, it goes beyond simple culling and becomes a matter of survival that leads to complete extinction. The Tower of Babel episode in the Old Testament is also a narrative that reveals human greed, pride, and competitiveness that cannot be stopped even if we know it.

This structural situation in which the brakes cannot be stepped on due to the limitations of human instinct has been conceptualized in various terms in social science, including game theory. ‘Prisoner’s Dilemma’ and ‘Chicken Game’ are representative examples. The Prisoner's Dilemma is the concept that two entities can achieve the best results if they cooperate with each other, but in the end, if they each pursue their own interests by distrusting the other, they will both suffer the worst results. In the game of chicken, the one who turns the steering wheel first among the two drivers charging head-on at each other loses. However, if the other person does not avoid it, it is an extreme confrontation that will eventually lead to catastrophe for both of them. There is a saying, ‘It’s like riding on the back of a running tiger.’ It is dangerous because the moment you get off your back, you will be eaten by the tiger, but you have no choice but to keep running.

Competition among global AI companies is exactly like this. On the surface, they say they should control the pace, but in reality, they cannot trust each other. As speed is the key to new technology, once a company is ahead, it must catch up quickly, but sometimes, once a company falls behind, it becomes impossible to catch up, let alone overtake. I am even more nervous because I don't know which company is developing new technology in complete secrecy. Co-extinction due to AI is an ambiguous problem that all of humanity will face in the distant future, and the survival of companies and individuals depends on whether or not we lead the AI ​​field. The reason humans have chosen the latter in these issues is an instinct that cannot be judged solely by moral standards.

The same goes for the AI ​​development competition between the United States and China, which are the two major AI powers and are currently competing for global hegemony. The Winner-Takes-All concept is recognized as a truth in the cutting-edge new technology industry, and this phenomenon is even more noticeable in inter-state relations. Between the US and China, the one who loses in the AI ​​competition may not only fall to second place but also face the risk of extinction. That is the fate of the empire. In particular, the fact that the future war is expected to move towards an 'AI unmanned war' is a key point where both countries can never give up on AI competition.

In fact, the leaders of both the US and China are dismissing the theory of AI speed control raised by academia and industry. This is because the decisive blow that will determine the struggle for hegemony between the two countries is in the field of AI. The Trump administration approaches it from the perspective of maintaining security and hegemony, putting forward a crisis theory that if the pace of AI development is slowed, it will be overtaken by China and lead to national failure. China goes one step further and adds conspiracy theories. He unwaveringly reaffirmed his policy to advance AI, saying that Western organizations' fear marketing and competition-suppressing tactics are hidden behind AI developers and scholars' speed control theory.

We know from experience that international law or international treaties can become a piece of paper at any time depending on the interests of the powers. Likewise, the AI ​​speed control theory is meaningless if the powerful and industry cores do not agree, even if everyone agrees with its purpose. Even though the fiction that AI caused nuclear war and destroyed humanity in the movie 'Terminator' is increasingly likely to become reality, it is not easy to control AI development. I don't know if there is an entity that can actually stop the US-China hegemony struggle and the development competition between AI companies, but unfortunately, there is no entity with such power yet in reality.