In 2017, China saw ALPHAGO, an artificial intelligence program supported by Google, broke Chinese geniuses in a complex board game, so he saw A Kei and shock. Similarly, the decisive loss to a foreign computer program, which had defeated Korean players, was a kind of Sputnik’s moment for China.
That year, Chinese officials made a bold plan to lead the world with AI by 2030, and pledged billions of companies and researchers focusing on technology. From this enthusiasm, by creating a strong AI model with much less money than the experts thought, DeepSeek, a Chinese emerging company that overturned the scenery of technology, has appeared.
Deepseek is private and has no clear support, but its success embodies the ambition of Xi Jinping, a Chinese top leader who recommends his own country to “occupy the commander.” Mr. XI hopes that the Chinese economy is driven by the most advanced technologies such as AI, Super Computing, and Green Energy instead of old growth engines such as real estate and inexpensive exports of debt fuel fuel. 。
For Xi Jinping, this moment is helpful to dent the dominant aura that has been held in AI, an important field in a rivalry of a superpower that the United States is intense. China has gained a status as a compassionate global partner in developing countries and shares the know -how. Mr. XI says AI must not be a “rich country and a wealthy game”.
At present, DeepSeek has indicated that China may be cheaper and easier for anyone to access AI. But the problem is how the ruling Communist Party manages the rise of technology, which can threaten its profits, and to grasp power.
AI’s Chinese regulations have changed their strength over many years, depending on where the country evaluates its strengths and weaknesses. If the Chinese government was worried that the United States would have been delayed in 2022 after the start of Openai’s ChatGpt, more hand -off -a -off is needed to enable ventures such as Deepseek.
Now that the pendulum shakes, Matt Seahan said that the confidence in the industry may be proven to be a “double -edged sword.”
“The core instinct is heading for control,” said Seahan. “As they regain their confidence in China’s AI ability, they may have a hard time to resist the urge to take more practical approaches to these companies.”
To emphasize the possibility, Liang Wenfeng, the founder of DeepSeek, has released the latest and most powerful AI model known as R1, and on January 20 with Prime Minister LI Qiang. I was invited to.
Liang attendees were even more surprising, given that DeepSeek was not considered one of the so -called AI tigers in China. The distinction is made for famous companies such as ZHIPU AI. This is a Beijing -based emerging company that has been invested in a considerable state.
DeepSeek is not a stranger to the urge of the party who wants to interfere. It may have played an inadvertent role in its final success. DeepSeek originally trained AI models to bet on the stock market in China. However, when the regulatory authorities targeted such actions, they proceeded with AI in 2023 to comply with China’s industrial policy.
Later, despite the fact that China has been using more advanced computer chips that are difficult to acquire, the world surprised the world by comparing the performance of American competitors. It is a technical feat that could not be used until recently. At home, Chinese commentators maintained DeepSeek’s achievements as evidence that US restrictions on exporting AI chips to China would ultimately waste (the founders of the company are such restrictions. )
Even the recent Openai claims that Deepseek had inappropriate data to build a model, has not stopped Chinese fans who have blamed San Francisco rumors.
“The US technical sanctions in China have no choice but to develop China,” said San Chengao, a diplomatic relations specialist in Tisin University in Beijing, reflecting popular emotions in China. I am. “We can rely on ourselves.”
AI has a special place in the vision of Xi Jin -kai on the rise of China, and it may support that country overcoming many of the largest tasks, such as reducing labor. China uses face recognition and algorithm to charge people and charge the ability to eliminate opposition. This technology also considers China’s military modernization using an autonomous weapon system. Battlefield strategy。
The development of Deepseek may lead to China’s geopolitical goals. DeepSeek uses an open source model. In other words, unlike a major American company that uses more expensive software, everyone can look under its hood and use technology.
“The low cost and open source nature of DeepSeek’s models are strengthening the Chinese government’s story that developing countries are the place where AI solutions are seeking AI solutions,” SHEEEHAN said.
How China grows in the AI global stage may depend on how the government will ultimately decide. Balance regulation There is freedom that companies and researchers need to do the most advanced work that can compete with the United States.
Some analysts, such as Gregory C. Allen, a researcher at the Strategy and International Research Center, are formerly American Defense in charge of the United States, and there is a high possibility that there is no limit to the development of AI for the Chinese army. I mentioned.
“The only thing that hinders them is the performance,” said Allen, who talked to the members of the People’s Liberation Army, who is responsible for evaluating AI’s risks in his previous job.
The same does not apply to regulating AI in the private sector. The landscape there is determined by the competitive priority of Chinese regulatory organizations, and each has avoided the technologies that many people in the world have not yet fully understood.
It is clear that the more technologies are used, the more they want to control them. In 2023, only a few months after ChatGpt caused an investment frenzy over artificial intelligence, China issued rules to control Chinese chatbots. Users are required to reflect the “core value of socialism” and avoid information that damages “national power”.
In the case of DeepSeek chatbots, this led to a troublesome reaction to seemingly good questions such as “Who is Xi Jinping?” Researchers who test their abilities have discovered that bots will provide an answer to spread the fake information campaign for Chinese propaganda and parrots.
Some concerns are essentially more and more existing. The growth of scholars’ chorus has issued alarms on potentially catastrophic consequences of losing human control against AI.
In those voices, the chief is Andrew YAO, a giant of the AI of Twin Island University and a winner of the Turing Award. His influence helped China to establish a global AI governance initiative introduced by XI in 2023, and was always called for maintaining AI under human management. 。 Last year, the government also called for “AI Governance based on human decision -making and directors”.
Eventually, China’s AI, Baras Halitas, an AI policy expert in the Washington Sub -Research Center, which is a Washington think tank, that the AI in China may only move forward. I mentioned.
“The need to comply with excessive regulations and” core socialist value “can cause the potential of AI to castrate,” said Harithas.
