[Into the World of AI] "Calling for Collective Slowdown, but Who Will Actually Stop First?"... Relentless AI Race Despite Speed Control Argument

China Accelerates the Race... Huawei to Launch AI Chip Early

Opposition to Joint Oversight Body in the U.S.

South Korea: "Speed Control Differs for Leaders and Followers"

Questions are being raised regarding the effectiveness of the so-called “AI slowdown” argument, proposed by leading U.S. artificial intelligence (AI) companies such as Anthropic and OpenAI, which calls for securing time to control technology-related risks. This skepticism is intensifying as China is accelerating its technological catch-up, and even within the U.S., disagreements persist over the establishment of a joint supervisory framework.

Yonhap News Agency

According to industry sources on September 19, Huawei plans to release its next-generation AI chip, the “Ascend 960DT,” in the first quarter of next year—three quarters ahead of its original schedule. The move comes as the company seeks to reduce its dependence on Nvidia and build a domestic AI computing base amid the U.S. government’s advanced semiconductor export restrictions.

On September 17, at “Huawei Connect” held in Shanghai, Huawei’s rotating chairman Xu Zhijun stated, “Chinese companies have not yet experienced the safety risks associated with advanced models that concern U.S. AI firms.” He continued, “Chinese AI companies need to further accelerate development so that they can reach a level where such risks are actually encountered and manageable.”

Indeed, in China, the AI slowdown argument is viewed as an attempt by leading U.S. companies to maintain their technological edge. China is also working to establish national standards for AI agent safety, but maintains a fundamentally aggressive approach to technology development. While acknowledging the necessity of ensuring safety, there is pronounced caution against slowing its pace of technological catch-up.

Diverging Opinions in the U.S.: "AI Speed Control Is an Individual Company’s Responsibility"

Yonhap News

There is no consensus on the AI slowdown argument in the U.S. either. According to the Wall Street Journal (WSJ), Meta CEO Mark Zuckerberg, Nvidia CEO Jensen Huang, and Tesla CEO Elon Musk last month expressed their opposition to the establishment of an AI regulatory body to U.S. President Donald Trump. The organization, proposed by Demis Hassabis, head of Google DeepMind, was envisioned to oversee the AI industry and establish standards in a manner similar to the U.S. Financial Industry Regulatory Authority (FINRA). These three leaders reportedly feared that such a body could concentrate authority in the hands of leading AI companies such as OpenAI, Anthropic, and Google.

They maintain that the responsibility for managing the pace of AI development belongs to individual companies, drawing a clear line against any industry-wide blanket slowdown. Jensen Huang, CEO of Nvidia, said during an AI-related meeting in the UK on September 17, “If you’re not ready, delay your launch.” He emphasized that companies need to rigorously ensure preparedness before releasing products.

U.S. President Donald Trump opposed restrictions, citing competition with China. On September 14, he asserted that the U.S. already has mechanisms to supervise AI firms and penalize illegal activity, and rejected calls for additional regulations. He argued that delaying AI and data center development could work in favor of competitor nations. This illustrates the collision between concerns about potential technological risks from private companies and the logic of securing national leadership in the sector.

The next-generation AI research efforts of these companies are also ongoing. Anthropic, which raised the AI speed control argument, announced on September 17 that Claude is spearheading 26% of its internal AI research and development work—a substantial increase from 1% in March. In August, the company reported that about 30,000 AI agents were simultaneously conducting research and engineering tasks on its internal platform.

South Korea: "No Time to Slow Down... Maintaining Speed to Catch Up"

The South Korean government has also stated it must maintain its development speed to catch up in the technology race. On September 16, Baek Kyunghoon, Deputy Prime Minister and Minister for Science and ICT, said on social media, “We currently have no room to slow down AI.” His argument is that it is fundamentally different for countries leading the way to control their speed, versus nations that must accelerate and create new markets as they catch up technologically.

This content was produced with the assistance of AI translation services.