With a liberal perspective, creativity is associated with personal freedoms because it allows its owners to pursue strange and unfamiliar ideas comfortably and without restrictions. In the next report from the American “Foreign Affairs” magazine, authors Carl Benedict Frey and Michael Osborne demonstrate the validity of this condition from a historical perspective.
In the past, it was widely expected that Japan would obscure the sun of the United States as a world leader in technology. In 1988, New York Times correspondent David Sanger spoke of a group of computer experts in the United States held a session to discuss Japanese technological development. When the group evaluated the new generation of computers coming from Japan, Sanger wrote: “Any illusions that America maintains a comfortable precedence. You are gone. “
Replace “computers” with “artificial intelligence”, and “Japan” with “China,” and it will look as if the article was written today. In a book titled “Artificial Intelligence Great Powers: China, Silicon Valley, and the New World Order,” which has become, unsurprisingly, a bestseller in no time, Kai Fu Lee, Google’s president in China, argues that the data treasure is unrivaled and the culture of copying is unrivaled. And the government’s strict government commitment to artificial intelligence, which China holds, gives it a great precedent for the United States. And more recently, Harvard University political scientist Graham Allison argued that China’s embrace of what most Americans see as a terrifying watchdog gives it a major data advantage over the United States.
As two researchers studying the applications of artificial intelligence and its implications, we respectfully disagree with them. And if China has any preponderance to replace the United States in the field of artificial intelligence, it will not go beyond that of Japan in the domination of computers in the eighties of the last century. Although China is rich in data, and although it has excelled in improving technologies invented in other regions, many obstacles prevent it from transforming into a new center in the major breakthroughs that AI badly needs.
China has made international headlines by effectively using its surveillance techniques to track down contacts in its response to “Covid-19, the disease caused by the emerging corona virus. However, the priority of the alleged data for the country is much greater than its actual size, and one of the reasons for this is that the data is associated with every field.” Separately, and that it does not usually solve more than the problems for which it was collected, and China’s excesses in the issue of privacy enable it to spy on its citizens, no other. That is because the abundance of surveillance data does not give China a preference in the application of artificial intelligence to areas such as drug discoveries or autonomous cars Driving, for example.
The puzzle of AI is not the amount of data accessed by the algorithms, but rather the effectiveness of its learning from that data. Even with huge amounts of data, AI systems are lured into making mistakes. Chretzian Segedi, a Google researcher and his colleagues, demonstrated this point by deceiving algorithms that once confidently and accurately sorted pictures of dogs and school buses separately, as the researchers manipulated pixelated images in a manner that made them difficult to monitor with the naked eye, and this led the algorithms to Classify both dogs and buses as ostriches. Artificial intelligence algorithms are often able to identify objects, but they lack any conceptual understanding of the relationships between those objects or their distinctive properties, in line with what we were warned by Joshua Bingo, a deep learning researcher.  When he said: “We cannot realistically label everything in the world and explain every detail with extreme precision for the sake of a computer.”
Many people like to think of China as “Saudi data.” But if the data is the new oil, then it could be a damned natural resource for China. For example, in the early 1920s, the electric car looked more promising than the gasoline-powered cars. But the huge oil discoveries, among others, have tipped the scales in favor of the internal combustion engine. A century later, we are trying to return to electric cars. And focusing on data-hungry AI applications could lead to similar retention in the wrong kind of AI.
We’ve seen this movie before. In the 1980s, huge promises and a massive focus on symbolic artificial intelligence led to massive media funding and intimidation, and that meant draining funding for “deep learning”. But deep learning has its share of problems that have led companies recently to focus on solving easier AI problems, such as sorting pictures of cats and dogs, where data is abundant, and this approach alone could reach a dead end that introduces AI in another winter.
Data efficiency is the sacred torch for advances in artificial intelligence, the main reason why most people associate the steam engine with James Watt instead of Thomas Newcomen, who developed a coal-fired steam engine decades before him, is that separating a watt of a capacitor first made that technology more energy efficient. The AI is still waiting for the moment to disconnect its capacitor. In order to win the Chinese goo game against Lee Sedol, the board strategy champion, Alpha Go DeepMind had to play against itself millions of times, and it had learned to move pieces more slowly than anyone would. Humans can benefit from data to an incredible extent, and recent breakthroughs in artificial intelligence are much smaller than that. The leadership of the United States or China in the field of artificial intelligence depends less on who owns the data than it does on who will be the first to innovate beyond this obstacle.
Those who warn about China’s stubborn advances in artificial intelligence are afraid because technology is inherently centralized, authoritarian governments will be better able to foster innovation than democratic governments, and that AI technology, in turn, will give preference to authoritarian governments. This fear calls for a belief that prevailed in the last century about electricity, and just like that, today’s belief is also misplaced.
In 1923, Charles Stements, whose work for the General Electric Corporation of America at the beginning of the 20th century made him a celebrity of that period, predicted that electricity would lead to the emergence of a more symbiotic society. Stimnitz argued, through erroneous reasoning, that developing a national electricity grid would lead to the emergence of socialism, because only a socialist system would be able to manage the interdependencies that would be required to advance toward a national grid.
Indeed, the “Rural Electricity Law” of 1936 provided funding to rural cooperatives that were neglected by the large private energy companies. But the shift to electric power has resulted in capital competition, in the form of factory trials. When engineers knew how to power each machine with its own electric motor, rather than relying on a centralized power supply, they could chain machines according to the natural flow of production, a breakthrough that led to mass production.
Thus experimentation and decentralized decision-making will be crucial if the world is to harvest AI crops. China does not have a preference in this regard, and the country’s recent boom in patent registration is often cited as evidence of its innovation, but the patent count is simply not a good way to measure creativity, and studies show that only 10% of patents represent 90% of the total value of patents around the world. This means that most patents have very little value. Patent citations are often a better indicator, and if we look at the 100 most cited patents since 2003, China was not responsible for any of them. Moreover, the leading AI companies in China, including Tencent, Alibaba and Baidu, are mere copies of Facebook, Amazon and Google but are tailored to the size of the Chinese market.
As Alexander Gershenkron, the late economic historian, pointed out, when a country lags behind in the technological race, the simulation and adoption of foreign technology can take a long time, and in general, the longer this country lags behind, the greater the role the state plays in leading industrial catch-up . Thanks to state investment in massive production of technology, the Soviet Union grew steadily during much of the Cold War’s lifetime, and so did Japan, South Korea, and Taiwan. Indeed, many researchers have credited the “Asian miracle” with catching up with the state-led industrialization. Although they managed to narrow something of the gap, these countries never managed to topple the United States. Unlike simulations, which can be planned and coordinated, creativity is a journey that delves into the unknown, if we reformulate the words of the economist and philosopher Friedrich von Hayek. Moving from simulation to creativity is difficult, and if it were easy, most countries would be able to innovate on the technological front.
By saying that China is unlikely to topple the United States from the throne of technological innovation, we do not mean by any means underestimating the tremendous economic achievements of China since Deng Xiaoping ascended to power in 1978. China is full of talent, but the fact remains that Chinese creativity remains so far a position. On the gradual improvement of technologies found elsewhere. Currently, Chinese companies lead the world in developing 5G networks, for example, but their work is based on many previous generations of communication technology. What Huawei is proving is that China has huge engineering capabilities, like Japan and certainly like the Soviet Union.
Artificial intelligence has not yet become a mature technology, and its continued advancement will depend on radical innovation on a number of fronts. Breakthroughs will happen as they always happen: by chance and recombination, in the interaction and exchange of ideas of inventors and entrepreneurs. As a strong and symbiotic structure, China has many advantages in terms of rapidly building infrastructure or massive rapid response to pandemics. But radical innovation is another matter, and throughout history, the most creative societies have been those that allow their citizens to pursue strange ideas. And as Joel Mokyr, a prominent economics historian, has argued, this is why the Industrial Revolution was taking place in the West rather than in China in the first place.
China’s efforts to restrict the flow of ideas online and elsewhere are likely to be an obstacle to creativity. Since September 2019, both China and Huawei have been introducing fundamental changes to the Internet infrastructure so that they contribute to global networks. If implemented, these changes are likely to further fragment the Internet and reduce the exposure of Chinese citizens to new ideas outside the country. This initiative records Beijing’s desire to preserve the political status quo, even if it means slower innovation and less dynamism.
Having said that, even the United States is not promised to win the race to dominate artificial intelligence, as China can still change its course, and the new restrictions imposed on immigrants by the Donald Trump administration in the United States may impede creativity in the United States, where it appears Research has shown that immigration has been a major driver of American creativity over the past 130 years. The Trump administration’s alleged plans to limit H-1B visas are particularly troubling in this regard, but while Trump may hold power for another term, Xi Jinping may rule indefinitely.
Under Chi, the Chinese Communist Party accelerated its efforts to penetrate the private business sector and consolidate political power. Ostensibly, state surveillance agencies and a controlled internet, along with a social credit system that promotes harmony and obedience, are unlikely to enhance creativity. Creativity is associated with breaking rules, not obeying them, and a study published recently in “Proceedings of the National Academy of Sciences” confirms that positive attitudes are Towards harmony and obedience is an indication of less revolutionary creativity.
Japan has failed to topple the United States, even without those restrictions on the flow of ideas and an authoritarian regime that promotes obedience. Consequently, the United States has crucial advantages that it should keep as the world’s leader in artificial intelligence. If it cedes this position to China, the reason will likely be Washington’s attempt to emulate the Chinese model by supporting nationalist figures rather than embracing the competition and dynamism that has made the United States the world’s technological front for more than a century.
- It is a new area of research that deals with finding theories and algorithms that allow a machine to learn on its own by simulating neurons in the human body.
This report is translated from Foreign Affairs and does not necessarily reflect the Maidan website.