As China's artificial intelligence (AI) ascendance shakes the global technology landscape, the U.S. big tech sector is moving to counter it by officially proposing the 'expulsion of Chinese AI.' Recently, as evaluations emerged declaring that the technologies of Chinese startups such as DeepSeek and Manus threaten OpenAI and Google, OpenAI urged the Trump administration to 'completely restrict the use of AI models developed in China.' This trend, coinciding with industry observations that AI has escalated into a 'national power struggle,' suggests that countries worldwide may enhance regulations surrounding AI related to security and technological hegemony.
◇ OpenAI designates 'DeepSeek' as a state-controlled organization
According to TechCrunch on the 13th (local time), OpenAI explicitly stated in a new policy proposal submitted to the U.S. government that DeepSeek is a 'state-controlled organization receiving government subsidies' and requested that the U.S. and 'allied countries (Tier 1 countries)' consider prohibiting AI models created by them.
The proposal included an explanation that 'DeepSeek is obligated under Chinese law to submit user data, making its AI models a serious risk factor for security and intellectual property (IP) protection.'
If OpenAI's policy proposal leads to actual institutionalization, it will become difficult for corporations or research institutes in the U.S. and allied countries to utilize Chinese models.
OpenAI has previously raised issues regarding DeepSeek's unauthorized 'knowledge distilling' of products such as the GPT series, but it is unusual for the company to demand a complete ban by citing 'direct ties to the Chinese government' this time.
DeepSeek has not provided an official stance on this matter. While it is true that DeepSeek has supported the state-led 'AI ascent' initiative, the evidence of direct control by the Chinese government has not yet been confirmed.
OpenAI's mention of expelling Chinese models symbolically reflects the recently accelerated 'U.S.-China AI hegemony war.'
Google also stated, 'Chinese AI technology has not brought forth entirely new innovations,' while urging governments around the world to 'raise security and safety standards.' Recently, Alice Freund, head of Google AI and emerging technology policy, emphasized in an interview with ChosunBiz on the 13th, 'Regardless of which country develops AI, there must be thorough security and privacy standards.'
◇ Chinese AI simultaneously drives domestic stock market and overseas investment
DeepSeek attracted attention for demonstrating reasoning abilities comparable to OpenAI's GPT technology with its R1 model unveiled last January. Along with Manus and other Chinese startups, they have also surpassed American models in global benchmarks by emphasizing actual work automation capabilities such as financial analysis, report writing, and website production. It is noted that the Chinese AI industry stands out with a strategy of maximizing performance through new learning algorithms while lowering expenses by utilizing low-spec chips.
There is also an analysis suggesting that the stage where China's AI ascent can be controlled merely by 'security concerns' has already passed. Andrew Ng, a professor at Stanford University, noted in a recent interview with Chosunilbo, 'China is encouraging open source at the national level, and if the U.S. insists solely on closed models, there is a high possibility that China will dominate the global AI supply chain.'
Although the U.S. is strengthening controls aimed at key components such as semiconductors, there is a prospect that 'second and third DeepSeeks' may continue to emerge based on the Chinese government's large-scale investments and vast pool of talent.
On the 7th, China announced plans to concentrate on fostering advanced industries like AI, 6G (sixth-generation mobile communications), and robotics during the National People's Congress and the Chinese People's Political Consultative Conference. President Xi Jinping has begun to elevate the AI ascent by directly meeting with DeepSeek founder Liang Wengpeng, empowering the heads of tech corporations.
In China, there are evaluations stating, 'Startups leading AI models, humanoid robots, and autonomous driving are rapidly growing, concurrently driving both the domestic stock market and overseas investments.' Alibaba and Tencent are also competing fiercely, releasing their own super-large AI models one after another.
If the U.S. government seriously considers expanding export controls or restricting the use of Chinese AI models due to security threats from AI, there are predictions that the global industrial supply chain will be shaken even more.
Im Jong-in, a special cyber advisor to the President, stated, 'The ban on Chinese AI proposed by OpenAI closely resembles the solution expected by the Trump administration,' and noted, 'If the regulatory scope is expanded to include products equipped with Chinese AI, significant repercussions in the global market are expected.'
He added, 'If AI competitiveness is somewhat hindered in sectors like home appliances, robotics, and automobiles where China has rapidly grown, domestic corporations could benefit from this,' and expressed, 'It would be prudent for us to closely monitor the actual policy direction of the U.S. after conducting thorough technical and legal reviews.'