Chinese AI Is Catching Up To The US Faster Than Before: Why It Matters
China's latest AI models are rapidly closing the gap with America's best.
National · 35 updates
China's latest AI models are rapidly closing the gap with America's best.
Kimi K3 hasn't dethroned OpenAI. But it has done something arguably more important.
American labs’ focus on closed, tightly controlled systems has coincided with a period in which one of them, Anthropic, was actively fighting its own government over access to its best models
Two Chinese companies, Huawei and Xiaohongshu, said their AI models achieved 100 percent at IMO.
AI rivalry deepens as US accuses Chinese startup of model distillation
Chinese AI firm Moonshot made headlines worldwide last week with the release of its open-weight, 2.8 trillion-parameter Kimi K3 model
US President Donald Trump said on Thursday that he will discuss artificial intelligence with Chinese President Xi Jinping when he visits the United States on September 24.
Anthropic CEO warns AI could aid bioweapons, calls China top threat
Meta CEO Mark Zuckerberg opposes banning Chinese AI models in the United States. He believes competition and identifying roadblocks are better solutions than outright bans. Zuckerberg also cautioned against American frontier labs potentially influencing policy to stifle domestic rivals. He reiterated support for open-source AI models, which are popular among Chinese developers. Meta is reportedly in talks to lease computing power to Anthropic for a significant sum.
Mark Zuckerberg says ban on Chinese AI could end up benefiting a handful of dominant AI companies.
Kevin O'Leary suggests China's growing power will unite North American economies. Canada offers minerals and energy, while the US provides scale and innovation. Mexico also plays a vital role in the regional economic landscape. Companies are increasingly using cheaper Chinese artificial intelligence models. This shift occurs as the race for technological dominance accelerates globally.
In practice, “open” comes with real friction as the weights are not conducive for self-hosting unless the user is well-resourced company
OpenAI cuts prices on smaller models as businesses scrutinize AI spend
China was never supposed to get this far, the US tried to ensure that through its sanctions. Beijing, however, had other ideas.
More than 80 Chinese papers and patents reviewed by Reuters show military-linked institutions using model distillation to build domestic AI systems for surveillance, cyber warfare and battlefield applications amid growing US-China technology tensions.
As Chinese AI models close the gap with their U.S. counterparts, military researchers are also examining distillation as a potential security risk
The issue has emerged as a major flashpoint ahead of US-China talks on AI governance and safety
Chinese military researchers are reportedly using advanced American AI models for training. They employ model distillation to transfer capabilities into smaller, locally controlled systems. This technique allows for defence applications, including drone operations and maritime analysis. OpenAI and Anthropic models have been identified in this research. Such methods are being explored amid US restrictions on advanced technology access.
Prominent figures in the US AI sector are advocating for startups to focus on serving the US government. This initiative arises amidst reports that Chinese companies are obtaining training data from American tech firms. Alarmingly, some US vendors cater to both government contracts and Chinese AI research labs, sparking national security concerns and jeopardizing US supremacy in AI.
The competition between the US and China in the realm of artificial intelligence has shifted focus towards harnessing model capabilities rather than merely their creation. AI model distillation enables smaller systems to mimic larger counterparts, which has sparked concerns over security. There are reports that Chinese military researchers are utilizing outputs from US AI models for defense purposes.
In a long essay published Monday about the future of artificial intelligence, Zuckerberg called for the U.S. to compete against China
Google and Meta both released new models this week, even as U.S. users wait for an open-weight model breakthrough
US last year launched Pax Silica initiative aimed at securing supply chains for AI models
Amid rising component prices caused by a worldwide shortage, the US government is urging Apple to discontinue sourcing memory chips from Chinese manufacturers. Allegedly, Apple has been assessing chips from CXMT and Yangtze Memory, however, US officials are pushing for a revitalized domestic supply chain. This call is echoed by US domestic chipmakers and senators who stress the importance of shunning Chinese memory suppliers in favor of local alternatives.
The draft letter, prepared by the State Department, is addressed to the 35 signatories of a U.S. “AI Opportunity Statement” signed in June
The United States is encouraging nations to choose their artificial intelligence development partners wisely. Countries that align themselves with China's technological model will face exclusion from the Pax Silica initiative. This strategy is designed to safeguard essential supply chains for AI technologies while restricting China's access to critical resources. To be deemed trustworthy partners, nations must align with a single technological framework.
The incident has prompted a Pentagon investigation