In 2017, China was in fear – and shock – as Alphago, the artificial intelligence program supported Google, beat Chinese penetration on a complex game on the board, GO. The decisive loss of a foreign computer program, which similarly ran into a player of South Korea, was a kind of sputnik for China.
That year Chinese officials exposed a brave plan Leading the world to AI by 2030, promising billions of companies and researchers focused on technology. From this Žarko came Deepseek, mostly an unknown Chinese start that In the midst of the technological landscape By creating a powerful AI model with far less money than experts considered possible.
Deepseek is privately, without obvious state aid, but his success embodies the ambitions of the best Chinese leader, Xi Jinping, who urged his country to “occupy command heights of” technology. Mr. Xi wants the Chinese economy to not be triggered by old growth engines such as debt and cheap exports, but the most advanced technologies such as AI, supercomputers and green energy.
For Mr. Xija, this moment helps a recessed aura of superiority The United States held in AI, A critical field in the fierce rivalry of the superpower. China acting as a well -intentioned global partner for developing countries, ready to share its knowledge, with Mr. Xi said that AI should not be “a game of rich countries and rich.”
Now Deepseek has shown that China could be possible for AI to make the cheaper and more affordable everyone. The question, however, is that the ruling communist party manages the increase in technology that could one day be so disturbing that it could endanger its interests – and its adherence to power.
The Chinese regulation of AI has varied over the years in intensity, depending on where the country assesses its strength and weakness. When the Chinese government took care of it, it lagged behind the United States in 2022 after the launch of Openi’s Chatgpt, it takes more approaches that ultimately enabled a venture such as Deepsek and to others.
Now that he has swung the other way in the other way, that confidence in the industry could be shown as “a double blade sword,” said Matt Sheehan, who studies the Chinese AI as an associate at the Carnegie Endowment for International Peace.
“The basic instincts of the party are according to control,” Mr. Sheehan said. “As confidence in Chinese AI capabilities, it may be difficult to resist the urge to approach these companies with a more practical approach.”
It was as if he had underlined the possibility, Deepseek’s founder, Liang Weenfeng, was called for a discussion with the Prime Minister Li Qiang on January 20, the same day when the company published its latest and most powerful AI model, known as R1.
The attendance of Mr. Liang was even more extraordinary given that Deepseek was not considered one of the so -called Chinese AI tigers. This difference is reserved for tall companies like Zhipe AI, a start-up based in Beijing, which received significant state investments.
Deepseek is not a stranger in the urgent party to interfere; This may have inadvertently played a role in his eventual success. Deepseek originally trained his AI models to bet on the Chinese stock market. But when regulators were aiming for such behavior, it turned the advanced AI in 2023 to align with Chinese industrial policy.
Then he stunned the world with the rival of his American competitors’ performance, despite using far less advanced computer chips that was hard to get China – Technological feat Until recently, this was not available. At home, Chinese commentators held Deepseek’s achievement because they are evidence that US export export limitations are ultimately in vain (although the founder of the company said such restrictions are the main concern).
Even recent accusations OpenSEEK that Deepseek has improperly collected his information for the construction of his models, did not distract their fans in China, who accuse San Francisco for spreading rumors.
“American technological sanctions on China left China without choice than to develop,” said Sun Chenghao, an external relations expert at Tsinghua University in Beijing, echoing popular feelings in China. “We can only rely on ourselves.”
AI has a special place in Mr. Xia’s vision about the rise of China, with the potential to help the Earth overcome many of its greatest challenges such as its festive workforce. China used to recognize and face algorithms Cover his ability to control His people and disagreement erupt. Technology is also conceived in Chinese military modernization with autonomic weapons and even Battlefield strategy.
Deepseek’s development could also improve Chinese geopolitical goals. Deepseek uses an open code model, which means that everyone can peek under the hood and use their own technology, unlike the leading US companies that use more expensive ownership software.
“The nature of Deepseek’s low cost and open code model tells the story of the Chinese government that China is a place in a place that can be developed in AI solutions,” said Mr. Sheehan.
How many cinema players would become on a global stage in AI, ultimately could depend on how the government decides Balance Regulations With the freedom that companies and researchers need to do a top job that allows them to compete with the United States.
Some analysts such as Gregory C. Allen, a researcher from the Center for Strategic and International Studies and a former US defense official, said that there are most likely no limitations in the development of AI when it comes to the Chinese army.
“The only thing that keeps them is a performance,” said Mr. Allen, who held talks with members of the National Liberation Army responsible for assessing the risk of AI in his former business
The same is not true of the regulation of AI in the private sector. The landscape there dictates competitive priorities of Chinese regulatory agencies, and each one feels around the technology that many in the world still do not fully understand.
Clearly, the widespread technology used, the more the party will want to divert it. 2023, just a few months after Chatgpt She initiated the anger of investment over artificial intelligence, China issued rules that are aimed at controlling what Chinese chatbots say to users, demanding that they reflect “socialist fundamental values” and avoid information that undermines “state power”.
In the case of Deepseek Chatbot, it led to Unpleasant answers On the seemingly benign questions such as: “Who is Xi Jinping?” Researchers who test his capabilities revealed that Bot gives answers to spread Chinese propaganda and even campaigns to disinform the parrot.
Some concerns are more of an existential nature. The growing choir of scientists sounded an alarm because of the potentially catastrophic consequences of loss of human control over AI
The main one among these votes was Andrew Yao, a DIV in AI at Tsinghua University and winner of the Turning Award, the Nobel Prize equivalent. His influence helped to determine what China calls Global AI Management Initiativewhich Mr. Xi introduced in 2023 and included an invitation to always keep Ai under human control. Last year, the Government also called on to improve the management of AI “based on human decision -making and surveillance.”
Finally, AI in China can only progress as much as the Government decides to alleviate these risks, said Barath Harrithas, AS expert AI at the Center for Strategic and International Studies, Washington Think Tank.
“Exaggeration and the need for adherence to” fundamental socialist values ”could risk placing the potential of AI,” said Mr. Harithas.
