US AI Industry Urges Against Broad Open-Weight Model Curbs Amid China Scrutiny
Several AI companies, including Nvidia and Mistral, have signed an open letter urging U.S. policymakers to avoid broad restrictions on open-weight AI models.
AI companies including Hugging Face, Meta, Microsoft, Mistral, and Nvidia have signed an open letter urging U.S. policymakers to avoid imposing broad, premature restrictions on open-weight AI models. The call comes as Washington debates its response to China's artificial intelligence activities, including allegations of intellectual property theft and model distillation.
While the letter does not directly mention China, it follows reports that the U.S. administration has considered banning Chinese open-weight models and potentially imposing sanctions. The White House has accused Moonshot AI of distilling data from Anthropic's Fable model to train its recently released Kimi K3 model.
The initiative appears aimed at preventing a complete ban on Chinese models and ensuring that the U.S. response to alleged data distillation does not extend to broader restrictions on open AI or common techniques like distillation. The companies emphasize that distillation is not inherently unlawful.
Industry players highlight that distillation, using one model's outputs to help train or improve another, is a widely used technique for model enhancement, evaluation, and validation. They compare this practice to the tradition within the open-source software movement of building upon existing technologies. The signatories hope policymakers will distinguish between deliberate misappropriation and advanced development methods.