矽谷与华府对于中国制人工智慧模型，特别是公开权重（open-weight）模型的扩散感到担忧。大型AI企业如 Anthropic 指控中国公司透过「模型蒸馏」窃取技术，并警告这些缺乏安全护栏的模型快速传播将带来严重安全风险。

然而，许多新创公司与知名创投强烈反对美国政府对这些模型实施禁令。他们认为，保持公开权重模型的自由存取能防止科技巨头垄断，并为资源有限的新创团队提供至关重要的平价工具，批评大型企业只是借由国家安全的名义来保护自身的商业利益。

尽管大型AI实验室不断强调开源模型的风险，但实际情况可能更为复杂。例如在最近一次骇客攻击中，专有模型的安全限制反而阻碍了调查，最终是依赖中国的公开权重模型才得以解决威胁，凸显出在商业竞争之外，开源技术在实际应用上的潜在价值。

Silicon Valley and Washington DC are growing concerned about the proliferation of Chinese-made AI models, particularly open-weight ones. Major AI companies like Anthropic have accused Chinese firms of stealing technology through model distillation, warning that the rapid diffusion of these models without safety guardrails poses serious security risks.

However, many startups and prominent venture capitalists strongly oppose a US government ban on these models. They argue that free access to open-weight models prevents monopolies by tech giants and provides crucial, affordable tools for capital-constrained startups, criticizing large corporations for using national security as a pretext to protect their own commercial interests.

Although major AI labs repeatedly emphasize the risks of open-source models, the reality may be more complex. For instance, in a recent hacking incident, the safety restrictions of proprietary models actually hindered investigation efforts, and a Chinese open-weight model was ultimately relied upon to resolve the threat, highlighting the potential practical value of open-source technology beyond commercial competition.