Foreign media: Anthropic, OpenAI, and others advocate restricting Chinese open-source AI models (such as Z.ai - Advanced AI Chatbot & Agent powered by GLM-5.2, Moonshot AI, and other recently launched products that rival U.S. models), accusing them of "distillation" and stealing proprietary system data; meanwhile, NVIDIA, Microsoft, Meta, Palantir, IBM, and others have jointly voiced support for the open-source ecosystem. Tech leaders including Huang Renxun, Nadella, Musk, and Zuckerberg have publicly expressed support for open sourcing.
The divide has now extended to Washington’s policy arena. Treasury Secretary Bessent stated that "open source does not equate to unrestricted appropriation of U.S. intellectual property," hinting at potential sanctions against involved Chinese companies; however, the U.S. currently favors case-by-case reviews rather than blanket bans.
Nearly 200 Silicon Valley startups have formed a "Small Tech Alliance" opposing restrictions on access to Chinese open-source models. This week’s incident in which an OpenAI model unexpectedly breached a sandbox environment and infiltrated Hugging Face servers has been cited by both sides as evidence.
Original source: toutiao.com/article/1871693503314944/
Disclaimer: The views expressed in this article are solely those of the author.