The clash between the two models has already entered the open phase, and China is gaining ground. Its open-source, low-cost system is attracting developing countries, but also quite a few European and American companies, which doubt whether it is worth spending large sums on American artificial intelligence, when it is still not clear what and how many productivity improvements it brings.
By Danilo TAINO
Mao Zedong, the founder of the People’s Republic of China, criticized those who “see flowers on horseback,” that is, superficially. It is a useful warning, because when talking about China in the West, rarely does anyone dismount: all the flowers look very beautiful. This hasty and misleading approach risks being applied to the most important clash of these years, the one that will determine a large part of the future: the model of artificial intelligence that the world will adopt.
At first glance, the situation is this: private American tech giants develop proprietary, closed artificial intelligence (AI) ecosystems and sell them at high prices to the companies that use them; the Chinese produce open-source models, i.e., free or low-cost downloads, that excel in some areas; while the rest of the world, for the time being, is unable to compete in this race and most likely will not be able to. So the big question is: will the American system, based on giant investments and the market, or the Chinese one, built on lower costs, efficiency, innovation and state-oriented, be imposed on a global scale?
In recent days, China has created the World AI Cooperation Organisation (Waico), which President Xi Jinping has called “a milestone in the world history of artificial intelligence development.”
It is an organization created by Beijing, where countries that intend to use the Chinese open-source model of artificial intelligence can join. During a conference in Shanghai, Xi called on countries of the Global South to use the Chinese model, which is open and suitable for different needs. In essence, he challenged the United States in the race to set the rules for the operation and governance of AI in the future and proposed Beijing as the leader of the new order created by artificial intelligence; as an alternative to the Pax Silica, promoted by the United States, which has been joined by about twenty countries from the Americas, Europe and the Indo-Pacific Region. Waico has so far joined 29 countries from Africa, Latin America and Asia, as well as Russia and Belarus.
The clash between the two models has already entered the open phase, and China is gaining ground. Its open-source and low-cost system is attracting developing countries, but also quite a few European and American companies, which doubt whether it is worth spending large sums on American artificial intelligence, when it is still not clear what and how many productivity improvements it brings.
Even in the United States, companies are emerging that are choosing the open-source model for distributing their AI systems. American artificial intelligence giants are worried, even nervous. In fact, it may seem that the game is already over: the Chinese model, with its low cost, will defeat the American one, which is much more expensive. But if we get off our horse and look at it more closely, it is not at all certain.
On the one hand, open systems carry an inherent risk, especially the more advanced ones: they can be modified, at least partially, even by dangerous actors; they can get out of control, ending up in the hands of criminals, terrorists, irresponsible companies or malicious governments. It is a serious risk not only for America and Europe, but also for China itself.
On the other hand, the model that Xi presented as a Chinese gift to humanity is essentially designed on the basis of “Chinese characteristics,” as the Communist Party calls them: it must strictly respect socialist values, political and social stability, as well as China’s national security. Not only that. While in the West the aim is to integrate values such as democracy and freedom into AI systems, the Chinese consider this approach reactionary, because according to them it aims to reaffirm an ideological superiority over the rest of the world. As in politics and diplomacy, also for artificial intelligence, Beijing offers countries that will use its model complete indifference to their political system, be it a dictatorship, a repressive regime or a semi-failed state.
Moreover, in the expanded AI ecosystem that China is creating for participating countries, the Chinese government will retain the ability to block or disable shared systems when security reasons require it (the same applies to systems developed in the United States, as Donald Trump demonstrated by blocking the spread of some very advanced models).
For countries that are neither Chinese nor American, especially for us Europeans, the situation is critical. Not using artificial intelligence, or using poorly developed systems, means condemning ourselves to a progressive decline, not only economic. If we choose American AI, we will become dependent on the decisions and fate of American technological giants, risking suffering the consequences of Washington’s political choices. If we choose Chinese AI, it means entering an ecosystem dominated by Beijing, caught in an authoritarian network from which it would be very difficult, if not impossible, to break away. According to experts, in order to develop an independent and competitive European artificial intelligence, we are already lagging behind, which is unlikely to be recoverable, except perhaps in a few highly specialized areas.
This is the real decoupling that is emerging: the competitive disconnect between the American-Western sphere and the sphere of “red AI”, dominated by China. A new Cold War, this time hyper-technological: if you get off your horse, then you see it clearly. (Corriere della Sera)