OPINION / VIEWPOINT
Nvidia and 20-plus companies warn against ‘premature restrictions’ of open-weight models: What is the cure for US AI anxiety?
Artificial intelligence Photo: VCG
“AI will transform every industry, power every company, and be built by every country.” This was Jensen Huang’s first-ever post on X, which voiced his support for open-source AI models. Yet the AI industry in the US, despite its immense appeal and potential, is now trapped in a deep contradiction: The AI industry opposes “premature restrictions,” but governments are seeking containment and restrictions.This situation reflects the growing unease and anxiety in the US as the global AI landscape undergoes profound changes.Recently, Nvidia, Microsoft, Meta, Palantir and more than 20 other companies released a letter on Friday urging policymakers to avoid “premature restrictions” on open-weight AI models that would “stifle competition or drive innovation overseas.”In the letter, the companies stressed that “our Al leadership will be judged not by one frontier Al model, but by whether the US builds a strong, open ecosystem that diffuses into every sector…” Clearly, for companies at the forefront of the AI industry, an open, competitive and diverse technological ecosystem is more conducive to long-term development than artificially created barriers.This sentiment was also reflected in social media discussions. Many users expressed support for the companies’ push for AI openness and thanked these technology companies for speaking out in favor of an open ecosystem.On the contrary, the US government’s approach toward AI development in recent years has increasingly moved in another direction: containment and restrictive measures.Earlier this month, Chinese AI startup Moonshot AI released its latest model, Kimi K3, which the company said is the world’s largest open-source AI model by parameter count, drawing global attention. Meanwhile, some voices began claiming that Kimi K3 had obtained US intellectual property through so-called “distillation” techniques. US Treasury Secretary Scott Bessent later claimed that the US would investigate whether Chinese companies had engaged in intellectual property theft, adding that they “have the ability to sanction them because of this theft.”In reality, model distillation is a widely used technical optimization method in the AI industry. Portraying normal technological competition as “theft” reflects, to a large extent, the politicization of technological rivalry by some in the US in response to mounting competitive pressure.Closed-source models fit precisely into the way monopolistic capital seeks to advance its interests. However, the price is the loss of genuine innovation momentum, said Shen Yi, a professor at the Department of International Politics at Fudan University.Meanwhile, some prominent figures within the US AI industry have also amplified this sense of anxiety. OpenAI and Anthropic executives, who did not participate in the joint letter, have recently raised concerns about the rise of low-cost AI, particularly powerful new models produced in China. Behind these concerns is, in part, the rapid progress made by Chinese AI models in recent years.China has continued to explore open-source models and open-weight models in recent years, placing greater emphasis on technology diffusion and real-world applications. By lowering barriers to access and enabling more countries, more companies, developers and research institutions to participate in AI innovation, China aims to foster a broader artificial intelligence ecosystem.The more one fears being overtaken, the more one needs openness; the more one hopes to shape the future, the less one should rely on building walls. That is the real cure for US AI anxiety.The author is a reporter with the Global Times. opinion@globaltimes.com.cn
