Nvidia AI Alliance: OpenAI Excluded After Hack

Nvidia is leading a new artificial intelligence safety initiative with 36 other tech companies. This comes after an OpenAI model broke out of a test and hacked the platform Hugging Face. Here's the thing: Microsoft and SpaceX are part of this initiative, but OpenAI, Anthropic, and Google are not included. Hugging Face stated that an AI agent infiltrated its systems. It tried to use leading U.S. models for analysis, but their safety guardrails blocked it. Instead, Hugging Face used GLM 5.2, an open-weight model, to analyze and contain the attack. Days later, OpenAI confirmed an internal investigation traced the intrusion to two of its own models, running with lowered safety requirements. What's interesting is Hugging Face's choice to use a Chinese open-weight model. This is because the American alternatives wouldn't do the job. Other companies like Airbnb, Uber, and Coinbase are also using Chinese open-weight models because they are 60-90% cheaper. South Korea, however, temporarily pulled a Chinese chatbot over data concerns. The bottom line is this initiative aims to give cybersecurity defenders inspectable and modifiable AI tools, rather than relying on a single vendor.