Tech giants are responding to a major cyberattack by creating an alliance to strengthen the security of open AI models.
Nvidia, alongside dozens of tech giants, has launched an AI security initiative focused on open models. This announcement comes as the consequences of a cyberattack involving a rogue OpenAI model continue to be felt across the global technology ecosystem.
A cyberattack revealing the limits of closed models
The revelation of a computer attack has highlighted the weaknesses of closed AI models. The target of the attack, the startup Hugging Face, was unable to use major US templates to defend itself because the guardrails did not distinguish between attacker and defender. Faced with this impasse, Hugging Face turned to a self-hosted Chinese open-weight model, which was not subject to the same restrictions.
This sequence highlighted a crucial point for the technology community: open models can be uploaded, modified, and self-hosted, unlike closed models — including cutting-edge systems built by Anthropic and OpenAI — which can only be accessed through specific infrastructure.
Faced with these challenges, Nvidia, Microsoft, SpaceX, Palantir and dozens of other technology companies from the United States and Europe have decided to unite. According to the official statement, “the Open Secure AI Alliance will work to patch and disclose vulnerabilities leveraging open technologies.”
The recent attack on Hugging Face served as a catalyst. “The recent security incident at Hugging Face serves as a clear reminder that defenders need open, cutting-edge agentic systems to defend themselves.” Nvidia also pointed out: “The OpenAI-Hugging Face incident showed a practical truth. When defenders cannot inspect, adapt and execute advanced AI on their own infrastructure, their ability to respond is limited precisely when speed is most crucial.”
Growing tensions around Chinese models
The technology sector opposes growing calls in Washington to limit access to Chinese AI models, the most successful of which are open. Their designers are accused of campaigns to extract information from the systems of American rivals, called “distillation”. Last week, Treasury Secretary Scott Bessent threatened sanctions against Chinese companies that carry out distillation attacks on U.S. companies.
Chris McGuire, senior fellow for China and emerging technologies at the Council on Foreign Relations, told CNBC: “There is a real possibility that the US government will impose restrictions on Chinese models.” He said this could include a ban on transactions involving the models, such as purchasing tokens through an API or hosting the model on the cloud by U.S. companies and charging clients for inference. He also specifies: “In Washington, it is not a debate on open source versus closed source, it is a debate on the tolerance or not of the theft of Chinese intellectual property”.
The majority of the most successful open source models are built by Chinese companies, raising concerns about potential restrictions. Last week, Nvidia, Microsoft, Meta, Palantir and more than 20 other companies released a letter urging policymakers to avoid “premature restrictions” on open-weight AI models that would “stifle competition or push innovation overseas.”
