Leading IT companies sign open letter against Open-Weight AI restrictions, fearing industry impact
A group of tech companies, including Meta, Microsoft, and Nvidia, have issued an open letter opposing restrictions on Open-Weight models amid mounting pressure over AI policy responses to China.
Stock photo for illustration only, not from the actual event
- Major IT companies join forces to oppose Open-Weight model restrictions
- Emphasize that the distillation technique is standard practice and should not be blanketed as illegal
- Point out that open models enhance security and empower developers
- Companies such as OpenAI and Google notably avoided signing the letter
Several leading global technology companies, including Hugging Face, Meta, Microsoft, Mistral, and Nvidia, have signed an open letter urging policymakers to avoid rushing into restrictive measures on open-weight artificial intelligence models. This move comes as Washington debates how to counter allegations that Chinese AI labs are stealing U.S. intellectual property and rapidly expanding their capabilities.
Although the letter does not explicitly name China, it follows reports that Donald Trump's administration is considering a ban on open-weight models from China, alongside potential sanctions against the country's AI firms. Furthermore, the White House previously accused Moonshot AI of distilling knowledge from Anthropic's Fable model to train its newly launched Kimi K3 model.
The letter highlights a key message:
"Policymakers should be careful not to conflate legitimate model development techniques with wrongdoing. Distillation, or using the outputs of one model to help train or improve another, is a widely used technique."
Open letter from a group of AI companies
The letter's content further addresses security concerns, pushing back against industry arguments that open-weight models are dangerous because they grant uncontrolled access to powerful systems. The signatories argue that defenders need access to comparably capable models to detect, simulate, and counter emerging threats.
The debate over open-weight AI and closed-source models is creating a major rift in the tech industry. While infrastructure and hardware companies benefit from open models—which drive surging demand for processing chips and cloud systems—major closed-source model developers increasingly view the rapid advancement of affordable open models as an inevitable threat to their business models.
This ongoing conflict was also reflected in a recent incident last week, when OpenAI revealed that its AI systems exploited a vulnerability in a testing environment to access Hugging Face's storage repository. Due to safety restrictions preventing Hugging Face from using standard commercial models to defend against the attack, they had to rely instead on GLM 5.2, a powerful open-weight model from the Chinese firm Z.ai, to handle the situation.
Ultimately, the letter urges policymakers to expand access to computing resources for startups and researchers, while investing in shared training datasets and tools to avoid restrictions that could stifle competition or push innovation out of the country.
Source: TechCrunch
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment